IF YOU WOULD LIKE TO GET AN ACCOUNT, please write an
email to Administrator. User accounts are meant only to access repo
and report issues and/or generate pull requests.
This is a purpose-specific Git hosting for
BaseALT
projects. Thank you for your understanding!
Только зарегистрированные пользователи имеют доступ к сервису!
Для получения аккаунта, обратитесь к администратору.
With latest changes in the udev, some deprecated functions were removed
from libudev amongst which there was the "udev_get_dev_path" function
we used to compare a device directory used in udev and directore set in
libdevmapper. The "/dev" is hardcoded in udev now (udev version >= 183).
Amongst other changes and from packager's point of view, it's also
important to note that the libudev development library ("libudev-devel")
could now be a part of the systemd development library ("systemd-devel")
because of the udev + systemd merge.
Support has many limitations and lots of FIXMEs inside,
however it makes initial task when user creates a separate LV for
thin pool data and thin metadata already usable, so let's enable
it for testing.
Easiest API:
lvconvert --chunksize XX --thinpool data_lv metadata_lv
More functionality extensions will follow up.
TODO: Code needs some rework since a lot of same code is getting copied.
When mirrors are up-converted, a transient mirror layer is put in so that
only the new devices are sync'ed. That transient layer must carry the tags
of the original mirror LV, otherwise it will fail to activate when activation
is regulated by lvm.conf:activation/volume_list. The conversion would then
fail.
The fix is to do exactly the same thing that is being done for linear ->
mirror converting (lib/metadata/mirror.c:_init_mirror_log()). We copy the
tags temporarily for the new LV and remove them after the activation.
Snapshots of RAID logical volumes are allowed (including "raid1"). However,
snapshots of "mirror" logical volumes has been disallowed due to unsolvable
issues inherent to the design. The fact that mirroring (dm-raid1.c) must
stop all I/O as the result of a failure and wait for userspace intervention
can lead to a circular dependency if userspace is simultaneously waiting for
snapshots (on mirrors) to make an I/O update before proceeding.
Various snapshot on mirror tests have been removed as a result.
Looking at the code in cmirrord/local.c, we can see the various different
request types handled in different ways. Some information that is non-changing
does not need to go around the cluster and can be short-circuited. For
example, once the cluster mirror is in-sync, it is pointless to continue
sending that query around the cluster. We can save network bandwidth and reply
directly back to the kernel. When it comes to status information, there are
two types 'TABLE' and 'INFO'. The 'TABLE' information never changes and
belongs to the group of requests that can be safely short-circuited. The
'STATUS' information can change - and will change if a device fails. Thus it
cannot be short-circuited, but this is exactly what was found. The 'STATUS'
information request was being short-circuited and therefore never reporting the
failure condition to anyone other than the "server" that experienced it
directly.
If two devices in an array failed, it was previously impossible to replace
just one of them. This patch allows for the replacement of some, but perhaps
not all, failed devices.
Thanks to agk for providing the patch that prevents resume from attempting
(and then failing) to create error devices which already exist; having been
created by a corresponding suspend operation.
In some occasional case dmevent restart was experiencing problems
with obtaining pid lockfile. So this patch tries to send several more kill
message until daemon kills itself so there is would reponse.
With this small loop the restart seems to work reliable,
although the loopsize and usleep are just randomly picked for now.
Just to make it clearer since there is the "dmsetup info -c -o blkdevname"
as well that shows the "block device name for this mapping", having a
"BlkDevName" header on output.
It's a bit confusing then if the "dmsetup info -c -o devs_used,blkdevs_used"
is named with a plural "DevNames"/"BlkDevNames" but at the same time having
a totally different meaning than the singular form "BlkDevName".
DevNames --> DevNamesUsed
BlkDevNames --> BlkDevNamesUsed
...makes it much more comprehensible.
The 'mirror' segtype and 'raid1' segtype both set the 'MIRRORED' flag.
However, due to differences in the way these device-mapper targets behave
'mirror' must be suspended with the 'noflush' option and 'raid1' does not
have to be.
This patch ensures that when the 'MIRRORED' flag is checked to see if
'noflush' is needed that it does not also set it for 'raid1' by mistake.
The logic for resuming the original and newly split LVs was not properly
done to handle situations where anything but the last device in the array
was split. It did not take into account the possible name collisions that
might occur when the original LV undergoes the shifting and renaming of its
sub-LVs.
When resizing thin pool - we need to use strip info from _tdata volume.
In future more generic solution will be necessary once we start to support
lvconvert (resize of stacked devices and stay properly aligned).
For now we just allow striped or linear LV so this code will work.
When given lvresize new size - round upward for stripes - unless we use % and
we are at the border of free extents.
This patch is not a complete fix and few more cases will need special care.
Libudev does not provide transactions when querying udev database - once we
get the list of block devices (devices/obtain_device_list_from_udev=1) and
we iterate over the list to get more detailed information about device node
and symlink names used etc., the device could be removed just in between we
get the list and put a query for more info. In this case, libudev returns
NULL value as the device does not exist anymore.
Recently, we've added a warning message to reveal such situations. However,
this could be misleading if the device is not related to the LVM action
we're just processing - the non-related block device could be removed in
parallel and this is not an error but a possible and normal operation.
(N.B. This "missing info" should not happen when devices are related to
the LVM action we're just processing since all such processing should be
synchronized with udev and the udev db must always be in consistent state
after the sync point. But we can't filter this situation out from others,
non-related devices, so we have to lower the message verbosity here for a
general solution.)
various dmeventd plug-ins into a new function called 'dmeventd_lvm2_command',
but the new function did not strip off the "_mlog" extentions that the
mirror plug-in had been doing. This created bug 794904 - failure to replace
devices in a redundant log.
The test suite did catch this scenario because it performs repair tests (mainly)
through the CLI and not dmeventd. It's also not easy to test because the test
itself will hang if the bug is encountered.
When vg_read fails, it internally unlocks VG if it's been locked,
so in error path we should skip unlock_vg for this case.
(user would see ugly internal warning)
Add make help target.
Add LVM_TEST_PARALLEL to support parallel runs of tests
Work around the problem the dmsetup table/info may return error
by using dmtable and dminfo function that will use 'should'.
(Error happens when some concurently running process removes table
entry while dmsetup command resolves table entries inside the loop.)
Activation on remote node should be tried only if it is masked by tags
locally (like when hosttags enabled, IOW activate_lv_excl_local()
doesn't return error.)
Introduced change caused that lvchange -aey succeeded even if volume was
activated exclusively remotely.
Calling vgscan alone should reuse information from the lvmetad (if running).
The --cache option should initiate direct device scan and update lvmetad
appropriately (if running).
This is mainly for vgscan to behave consistently compared to pvscan.
locally or on more nodes while others are activated exclusively.
Current pvmove code can either use local mirror (for exclusive
activation) or cmirror (for clustered LVs).
Because the whole intenal pvmove LV is just segmented LV containing
segments of several top-level LVs, code cannot properly handle
situation if some segment need to be activated exclusively.
Previously, it wrongly activated exclusive LV on all nodes
(locing code allowed it) but now this is no lnger possible.
If there is exclusively activated LV, pvmove is only
possible if all affected LVs are aslo activated exclusively.
(Note that in non-exclusive mode pvmove still activates LVs
on other nodes during move.)
# lvchange -aly vg_test/lv1
# lvchange -aey vg_test/lv2
# pvmove -i 1 /dev/sdc
Error locking on node bar-01: Device or resource busy
Error locking on node bar-03: Volume is busy on another node
...
Failed to activate lv2
In this case we should allow to use local mirror, check for cmirror
should apply only for lvconvert/lvcreate.
Introduced in 2.02.86 by removing !(lv->status & ACTIVATE_EXCL).
(Partially workaround, it is minimalistic patch for now.)
Code adds better support for monitoring of thin pool devices.
update_pool_lv uses DMEVENTD_MONITOR_IGNORE to not manipulate with monitoring.
vgchange & lvchange are checking real thin pool device for existance
as we are using _tpool real device and visible LV pool device might not
be even active (_tpool is activated implicitely for any thin volume).
monitor_dev_for_events is another _lv_postorder like code it might be worth
to think about reusing it here - for now update the code to properly
monitory thin volume deps.
For unmonitoring add extra code to check the usage of thin pool - in case it's in use
unmonitoring of thin volume is skipped.
Make the teardown really usable - it will try down to remove all the left
devices even from previous test runs
(the only missing piece is probably proper mdadm teardown)
Add few more local vars
Try to setup PATH and LD_LIBRARY_PATH just once.
Try shorter sleeps.
Actually restart was failing for different reason - so pass in proper
location of dmeventd for restart from lvm command and avoid using
the one from /sbin location.
Update pv create test with "" around path.
There are kernel drivers (smblk) which set '-1' as their device major number.
This number is listed in /proc/devices then - but the kernel itself is using
just 12 bits - thus device is accessible via 4095 - there is posted patch
for 3.4 to fix this behavior (0 for auto allocation was mean to be used).
However to still allow using such devices with older kernels add some code
to use same behavior - so cut 12 bits from the major number from /proc/devices.
For now use log_warn() - maybe the severity of the message could be lowered
to just verbose level.
Indent
Shell improvements - use internal function for checks
Use PVs in "" (LV and VG cannot have spaces)
Several test very starting 'dmeventd' without annoucing
it via prepade_dmeventd.
Fix some of test actually.
Indent
Better shell usage
Function simplification
More usage of 'get' functions
Don't use valgrind tracing for check and get function (faster)
Update shell debugging (PS4, better stacktrace)
Support paths with spaces
Export SCRIPTNAME for external usage
Watch for dmeventd unexpectedly started during test
Indent
Add valgrind support:
env LVM_TEST_VALGRIND={0123} (the higher level, more commands tested)
env LVM_TEST_CLVMD=1 runs clvmd within valgrind.
env VALGRIND script name executed for each lvm command (def. is valg).
Smarted teardown - should minimize occurence of left dev entries
(using dmsetup remove -f only as last resort)
sort removed devices by open count before actual removal
Use "" around string that may contain spaces.
Set log/verbose and activation/retry_deactivation to defined value.
Remove debug.log after successful lvm command (easier to check output).
Currently we could not test special prefixes in our test suite.
As teardown will not find such device and basicaly busyloops here,
as at cannot remove such names.
This patch adds possibity to use:
vgcreate V_$vg1 $dev
Note: you still need to use $PREFIX somewhere in the name.
(And of course, it's really bad idea to use $PREFIX (=LVMTEST)
for normally used LVs)
The only purpose of this patch is to allow testing cluster with
special vg names that begins with V_ , P_....
Fix propagation of -e option - pass it via internal shell variable.
Fix parsing of /proc/mounts files (don't check for substrings).
as reported by O.Mangold with suggested patch:
https://www.redhat.com/archives/linux-lvm/2012-February/msg00030.html
Properly pass arguments with spaces ("$@")
Add validation for YES and EXTOFF variable content.
When down-converting a RAID1 device, it is the last device that is extracted
and removed when the user does not specify a particular device. However,
when a device is specified (and it is not the last), the device is removed and
the remaining sub-LVs are "shifted down" to fill the hole. This cause problems
when resuming the LV because if the shifted devices were resumed (and thus
renamed) before the sub-LV being extracted, there would be a name conflict.
The solution is to resume the extracted sub-LVs first so that they can be
properly renamed preventing a possible conflict.
This addresses bug 801967.
Update a way we handle option passing - so we now support path and options
with space inside.
Fix dm name usage for thin pools with '-' in name.
Use new lvm.conf option thin_check_options to pass in options as string array.
LISTEN_PID and LISTEN_FDS environment variables are defined only during systemd
"start" action. But we still need to know whether we're activated during
"reload" action as well - we use the reload action to call "dmeventd -R"/"lvmetad -R"
for statefull daemon restart. We can't use normal "restart" as that is simply
composed of "stop" and "start" and we would lose any state the daemon has.
Save some relocation entries and use directly char[].
Since we do not need yes more then 127 partitions per device, use just int8_t.
Move lvm_type_filter_destroy into local static function.
Support timestamping with harness - using VERBOSE=2
Fix also logging in several situation
(i.e. continue logging multiple test in VERBOSE mode,
do not coredump with empty output).
Never return unfinished toolcontext - since error path is hit on
various stages of initialization we cannot leave it partially uninitialized,
since we would need to spread many more test across the code for config_valid.
Instead return NULL and properly release udev library resources as well.
We can't use 'DM_SBIN_PATH'. This one is set only for DM devices but not
for all block devices - the pvscan is run on all relevant block devices!
This LVM_SBIN_PATH (as well as DM_SBIN_PATH) detection should be removed
eventually but for upstream solution, we still have to do that as there are
known cases where the binaries are put either in /sbin or /usr/sbin
(some installation systems, initrd systems etc.).
Fix regression in man page. The chunk size is in kilobyte units on command line
input though in the source code we work with sector size unit
so make it clear in the man page.
Update chunksize for thin pool in man page - it's max value is 1024M == 1G.
Fix warning range message to show proper max value.
If the lvcreate may decide some automagical values for a user,
try to keep the pool metadata size into 128MB range for optimal
perfomance (as suggested by Joe).
So if the pool metadata size and chunk_size were not specified,
try to select such values they would fit into 128MB size.
Use thin_dump --repair suggestion in log error message
and use just warning on deactivation path without repair info
(since node has been deactivated).
Also check whether there is not 16 args for thin_check configured.
Auto mode can't deal with multiple mangled names. We can do that while working
in hex mode, but in auto mode, this would lead to device name ambiguity.
Be more strict when unmangling names on ioctl return - require the name to be
properly mangled in 'auto' and 'hex' mode. There really should not be any
blacklisted character since the names should be renamed already (by means of
renaming it directly or running 'dmsetup mangle' for automatic rename).
Avoid using NULL pointers from udev. It seems like some older versions of udev
were improperly returning NULL in some case, so do not silently break here,
and give at least a warning to the user.
if the thin_check fail on thin pool - still return successful deactivation,
since lvremove would currently fail.
TODO: find some way to not run check with lvremove.
Test if thin_check is present in system and disable its use, when its missing.
Add testing for poolmetadatasize.
FIXME: Allocation policy for metadata pool might need some relaxing.
(For now it needs to put all block on one PV.)
Reduce disc excercise for some test and focus on LVM testing by
using smaller extent size.
Reduce number of teardown_devs calls and use vg/lvremove instead.
Don't sleep for seconds on pvmove.
FIXME: shell/lvconvert-mirror-basic.sh seems to need more checking.
Test fails for smalled extent size then 512k.
Add some hack math to allow 16GB devices to be passed as thinpool metadata.
Since kernel has put in limit to not allow which are just bigger then
some predefined constant in kernel but not matching 16GB so any device bigger
is rejected.
FIXME: Current code still might need more tweaks to be more generic.
Use libdm callback to execute thin_check before activation
thin pool and after deactivation as well.
Supporting thin_check_executable which may pass in extra options for
the tool.
Add 3rd daemon return state "unknown" for lookups that are carried out
successfully but don't find the item requested.
Avoid issuing error messages when it's expected that a device that's
being looked up in lvmetad might not be there.
Move the code for poolmetadatasize operation into one place.
Report override for minimum and maximum size.
Drop _read_thin_params function its error reporting is handled elsewhere.
If no size was give the later added minimal size check efectively
disable this code. Also the argument for size now must be kept
in sector_size, so adding division by SECTOR_SIZE (moved into
a const expression)
Hold global lock in pvscan --lvmetad. (This might need refinement.)
Add PV name to "PV gone" messages.
Adjust some log message severities. (More changes needed.)
F17 is getting rid of OpenAIS libraries (and checkpointing). While the
CPG stuff is staying, some if its constants are being removed. So, we
must adjust and use the remaining constants which the CPG constants were based on.
[~]# egrep 'CPG_DISPATCH_ALL|CPG_OK' /usr/include/*/*
corosync/corotypes.h:#define CPG_DISPATCH_ALL CS_DISPATCH_ALL
corosync/corotypes.h:#define CPG_OK CS_OK
Since lvm seems to call driver_version(NULL, 0) this would lead
to crash. Though the combination of the code is probably very hard to hit.
If the user doesn't supply version buffer, just skip printing to buffer.
If the thin pool has disabled zeroing (created with -Zn), we at least
clear initial 4KiB of such thin volume (provisions 1st block).
If lvcreate is executed with '-an' command will abort (same way like we for
normal LV - however for normal LV option -Zn may skip clearing completely,
for thin volumes this option is not supported (applies only for pools).
The OpenAIS checkpoint library is going away; therefore, cmirrord must
operate without it. The algorithms the handle the timing of when to send
a checkpoint, the determination of what to send, and which ongoing cluster
requests are relevent with respect to the checkpoints are unaffected. We
need only replace the functions that actually perform the storing/transmitting
and retrieving/receiving of the checkpoint data. Rather than store the
checkpoint data in an OpenAIS checkpoint file, we simply transmit it along
with the message that notifies the incoming node that the checkpoint is
ready.
pvcreate gives
WARNING: Ignoring unsupported value for metadata/pvmetadataignore.
It was warning if there is no config file entry instead of only if the node
exists but is empty.
Drop whole buffer clearing (most messages at <100 bytes).
Just make sure we have always \0 terminated string for strlen() operations.
(before for PIPE_BUF sized messages this was not set).
Addressing somewhat tricky bug here.
Since stdin,stdout,stderr were closed it's been occasionally possible to
see some unexpected messages to be flowing into a clvmd and generating some
randomly sized allocation of many megabytes. Since the message was not
being generated by standard send_message() construction, after some more
testing it apperead to be a debug log message - thus something has flown
to local socket opened on strandard out descriptor.
To fix the issue - use standard file descriptor duplication code for daemons.
For making easier debugging of polling daemon - developer might want to recompile
without modifition of standard file descriptors.
This could be seen as some sort of simple validation - it's not easy to
recognize a valid message for now - but we definitely do not want to
allocate a lot of megabytes in clvmd memory locked daemon when broken
message gets in.
Size of 8000 is just selected for now - possibly there could be much
lower value put in.
As API is passing structures by value, do not leave
the function which created buffer and keeps valid pointer
look like it would be some memory leak and move
free of buffer from inner function - makes more obvious,
how is the memory management handled.
Using report_type_t for bitmask is not correct, since we have not defined types
for all bit combinations - so switching to unsigned type, since values of
report_type_t enum are unsigned.
Make sure both hash tables are initialized before _read_sections() call.
Presents no functional change (since PV scan phase was not adding LV hashes),
but makes the code easier to handle mem failing case, and static analyzer is
hapier as well.
Adding at least stack traces with some FIXMEs for cases,
where we might want to do something cleaver - maybe fail command
or give user hints something is not going well ?
For remote_backup is stack probably 'good' enough for now.
We cannot do anything better here anyway - we are already in logging function,
so just ignore this issue here - it will most likely stop application later.
Code cannot proceed if oldname would be NULL.
Since lvmetad currently doesn't use logging mechanism of lvm to report
internal errors - stay with current code style of lvmetad which uses
plain asserts for cases like this.
Why using the order 69:
- Storage processing in general happens in 60-persistent-storage.rules,
including the blkid call that adds some usable information we can use
for filtering and speedup (these rules are part of upstream udev and
the order is preserved on most distros)
- There's still some other storage-related processing done after
60-persistent-storage.rules in general. These might add some detailed
storage-related information we might use to filter devices effectively
(e.g. MD udev rules, ...).
- We need lvmetad rules to be processed before any consumers can use the
output - so the metadata cache is ready soon enough (e.g. udisks rules).
- There's no official (upstream udev) document about assigning the order,
so this number is chosen in best belief it will suit all scenarios.
Should be faster then strncpy - since we could avoid clearing 4KB pages
with each strncpy(...,PATH_MAX).
Also it's easy to check whether string fit - and eventually avoid
to continue working we incomplete string.
If we have good enough glibc to return number of needed chars, do not
loop try to reach good size, but use this size directly for allocation,
saving also last strdup.
Since now we start with 16 bytes - skip buffer realloc for shorter string.
make devices invisible to lvm, but the behaviour of those is slightly different
than of actual missing devices. Running vgscan after re-enabling the device
triggers a metadata repair which is not done by vgremove -ff. This is not a
regression, merely an odd behaviour that has been around even before lvmetad.
The code fail to account for the case where we just need a single device
in a RAID 4/5/6 array. There is no good way to tell the allocation functions
that we don't need parity devices when we are allocating just a single device.
So, I've used a bit of a hack. If we are allocating an area_count that is <=
the parity count, then we can assume we are simply allocating a replacement
device (i.e. no need to include parity devices in the calculations). This
should make sense in most cases. If we need to allocate replacement devices
due to failure (or moving), we will never allocate more than the parity count;
or we would cause the array to become unusable. If we are creating a new device,
we should always create more stripes than parity devices.
/etc/tmpfiles.d directory holds configuration files for temporary/volatile
files and directories that should be automatically managed. For example,
if we have some parts of the fs hierarchy on tmpfs, we'd like to recreate
some files or directories on every boot so they're always prepared for use.
Systemd can read such configuration files. For now, the lock and run directory
are the ones that are most probably placed on tmpfs. If this is the case, we
can install the configuration by 'make install_tmpfiles_configuration'.
There were no messages printed upon completiion of RAID device replacement.
This could cause confusion/concern during automated recovery, because the
user sees the failure messages but no other messages indicating correction.
Built-in blkid is supported since udev v176 - set the UDEV_HAS_BUILTIN_BLKID
variable appropriately so we can use it in the rules to call the built-in
blkid conditionaly.
config tree per PV which is mostly provided by the client, so it can be used to
keep track of things like label_sector, PV format, mda count / offsets and so
on.
Read lvm.conf setting for monitoring for each command. So we should not
activate monitoring if the default compilation is set to monitor during
lvconvert commnads.
Patch also removes check for clustered VG and allows to disable monitoring
for clustered VG with the assumption, the problem with monitoring and dmeventd
flag passing for INGNORE is already fixed.
We don't have anything better yet...
The problems the watch rule caused when removing devices should be covered
now with the "retry remove" logic. It's also better to have this maintained
by us, rather than having this rule anywhere else without proper control.
Device-mapper in kernel uses '\' as escape character so it's better
to double it to avoid any confusion when using existing device names
with '\' in the table specification.
For example:
dmsetup create x --table "0 8 linear /dev/mapper/a\x20b 0"
should pass just fine now without a need to explicitly escape the '\' char
like this:
dmsetup create x --table "0 8 linear /dev/mapper/a\\x20b 0"
If dm_task_get_name or dm_task_get_names gets called, these will return
unmangled form of the names so the name mangling stays totally transparent
to any libdevmapper user (unless DM_STRING_MANGLING_NONE is used in which
case the name is not touched and it is is returned as it is in kernel).
For example:
dmsetup create "a b" - will create a\x20b device in kernel and so udev will
create /dev/mapper/a\x20b
dm_task_get_name/names will still return "a b"
In AUTO mode, the libdevmapper user can still query the device by using
the mangled ("a\x20b") or unmangled form of the name when calling dm_task_set_name.
If mangled name is provided, it's detected and the name is kept as it is.
If unmangled name is provided, it will be mangled. IOW in AUTO mode it's
totally transparent and it should not require any changes in the code
using libdevmapper.
However, any libdevmapper user must be aware of the fact that the mangled form
of the name appears in /dev/mapper (udev just can't deal with those blacklisted
characters).
- rename the hashes to be explicit about the mapping
- add VG/PV listing calls to the protocol
- cache slightly more of the per-PV state
- filter cached metadata
- compare the metadata upon metadata_update
dm_task_get_name_mangled will always return mangled form of the name while
the dm_task_get_name_unmangled will always return unmangled form of the name
irrespective of the global setting (dm_set/get_name_mangling_mode).
This is handy in situations where we need to detect whether the name is already
mangled or not. Also display functions make use of it.
Use the DEV_NAME macro to use the mangled form of the name if present,
use normal name otherwise (we store both forms - mangled and unmangled in
struct dm_task). Mangled form should be always preferred over unmangled
with the exception of the situations where we divide one task into several
others (like "create and load") - we need to avoid mangling the name twice
(because of multiple dm_task_set_name calls)!
If dm_task_set_name/newname is called, the name provided will be
automatically translated to correct encoded form with the hex enconding
so any character not on udev whitelist will be mangled with \xNN
format where NN is hex value of the character used.
By default, the name mangling mode used is the one set during
configure with the '--with-default-name-mangling' option.
This option configures the default name mangling mode used, one of:
AUTO, NONE and HEX.
The name mangling is primarily used to support udev character whitelist
(0-9, A-Z, a-z, #*-.:=@_) so any character that is not on udev whitelist
will get translated into an encoded form \xNN where NN is the hex value
of the character.
It was not possible to pass down the DM_[FORCE|NO]SYNC flags to
'dm_tree_node_add_raid_target'. This meant that converting to 'raid1' from
'mirror' would cause a full resync. (It also meant that '--nosync' was
ineffective when creating a 'raid1' LV.)
I've taken the 'reserved' parameter in 'dm_tree_node_add_raid_target' and
used it for the "flags" parameter. Now it is possible to pass the sync
flags and any other flags that may come up.
s/Issue/Use/, otherwise it is easy to misread "Issue" as "Issuing" - causing
the user confusion as to whether the action was performed automatically or
whether they need to issue the command.
'_lv_update_log_type' takes a lvconvert_params argument so that it can pass
down the user's preference of 'region_size' and allocation_policy. When
'mirror_remove_missing' was introduced (commit ID
95986e42a1) it didn't make sense to pass down
user preferences - so NULL was given instead. While it may never happen in
practice, static analysis reveals that this argument could be dereferenced.
So, if the user preferences were not passed in, glean the necessary fields
from what is already set in the LV.
Reported-by: Zdenek Kabelac <zkabelac@redhat.com>
Signed-off-by: Jonathan Brassow <jbrassow@redhat.com>
(Not updating WHATSNEW for this simple clean-up.)
Commit 02f6f4902f introduced a bug that caused
RAID devices to fail to activate if the device for a single sub-LV failed.
The special case of LVM mirror was handled, but not LVM RAID.
EXAMPLE:
[root@bp-01 ~]# devices vg
LV Copy% Devices
lv 100.00 lv_rimage_0(0),lv_rimage_1(0)
[lv_rimage_0] /dev/sde1(1)
[lv_rimage_1] /dev/sdh1(1)
[lv_rmeta_0] /dev/sde1(0)
[lv_rmeta_1] /dev/sdh1(0)
[root@bp-01 ~]# vgchange -an vg
0 logical volume(s) in volume group "vg" now active
[root@bp-01 ~]# off.sh sdh
Turning off sdh
[root@bp-01 ~]# vgchange -ay vg --partial
Partial mode. Incomplete logical volumes will be processed.
Couldn't find device with uuid fbI0YO-GX7x-firU-Vy5o-vzwx-vAKZ-feRxfF.
Cannot activate vg/lv_rimage_1: all segments missing.
0 logical volume(s) in volume group "vg" now active
AFTER this patch:
[root@bp-01 ~]# vgchange -ay vg --partial
Partial mode. Incomplete logical volumes will be processed.
Couldn't find device with uuid fbI0YO-GX7x-firU-Vy5o-vzwx-vAKZ-feRxfF.
1 logical volume(s) in volume group "vg" now active
[root@bp-01 ~]# devices vg
Couldn't find device with uuid fbI0YO-GX7x-firU-Vy5o-vzwx-vAKZ-feRxfF.
LV Copy% Devices
lv 100.00 lv_rimage_0(0),lv_rimage_1(0)
[lv_rimage_0] /dev/sde1(1)
[lv_rimage_1] unknown device(1)
[lv_rmeta_0] /dev/sde1(0)
[lv_rmeta_1] unknown device(0)
[root@bp-01 ~]# dmsetup table vg-lv; dmsetup status vg-lv
0 1024000 raid raid1 3 0 region_size 1024 2 253:2 253:3 - -
0 1024000 raid raid1 2 AD 1024000/1024000
No WHATSNEW update necessary because this is an intrarelease fix.
brassow
Move commod code to destroy orphan VG into free_orphan_vg() function.
Use orphan vgmem for creation of PV lists.
Remove some free_pv_fid() calls (FIXME: check all of them)
FIXME: Check whether we could merge release_vg back again for all VGs.
In case of zero bytes would be read from sysfs, it would store '\0' on
temp_buf[-1] address.
Simplify some buffer length calculation and use strcpy if we've just
checked string fits in give buffer.
Replace jump label error: with bad: commonly used in libdm.
Replace asserts with test for failing memory allocation.
Add at least stack traces.
Index counter starts from 1 (0 reserved for error), so replacing fingerprint.
label_exit() is called destroy_toolcontext() and we are now
using standard dm_list function for destroy, we have to make sure
dm_list gets initialized properly.
Tested condition has been already evaluated before
For strlen() code has already excluded <ID_LEN.
For repairing, already tested (!argc && !repairing) before.
As atoi may return negative value - test for both limits.
Test log_args for limits before calling alloca().
Code from dmeventd mirror plugin should probably share same code as
we have in mirrored.c.
Since the function dm_get_next_target() returns NULL as 'next' pointer
so it's not a 'real' error - set 0 to all parameters when NULL is
returned because of missing head.
i.e. one of use case::
do {
next = dm_get_next_target(dmt, next, &start, &length,
&target_type, ¶ms);
size += length;
} while (next);
Since the function dev_close() has code path, which really could close
file (for unlocked vg) and destroy dev handler, stay on safe side and move
the close few lines later, even our current use case shouldn't trigger
such scenario.
Count number of error and existing areas and if there is no existing area
for the LV avoid its activation.
Always disable partial activatio for thin volumes.
For mirrors currently put in hack to let it pass with a special name
since current mirror code needs to activate such LV during some operations.
For reading % of mapped size of thin volume use as origin for
old style snapshot '-real' device needs to be queried.
Fix log_error report given for lvs -a in this case.
Before removing thin pool LV always make sure, stacked message
for previous run are cleared - but allow to remove any
device that should have been created
(i.e. creation of snapshot failed - so the message for snapshot creation
may be replaced with delete message within unfinished transaction).
Also commit messages after lv remove - so free space is released in pool.
Extend the usage of origin_only flag to allow resume of thin pool LV
(when it's active) to pass only the messages.
origin_only flag will skip detection of already resumed tree for thin_pool,
so we do not need to suspend the tree and we just send messages.
Pass in the origin_only flag also for thin volumes - but curently the flag
is not used to its best.
FIXME: achieve the state where only thin volume snapshot origin is
suspended without its childrens - let's explore whether this may
happen automatically inside libdm (might be generic for other targets).
So the code would not need to annotate the node for this.
Extend lv_activate_opts with bool flag to know for which purpose
dtree is created - and add message only for activation tree
(since that's the only place that may send them).
Extend validation check for thin snapshot creation and test whether
active snapshot origin is suspended before its snapshot is created
(useful in recover scenarios) - in this case also detect, whether
transaction has been already completed and avoid such suspend check
failure in that case.
Add pool_has_message and use it in attach_pool_message.
Also update header to make more obvious which segment type is
expected as parameter.
Rename 'read_only' to 'no_update' (no auto update transaction_id)
to better fit how it's used.
Fix problem when there was only one stacked message replaced with delete
message that caused unwanted transaction_id increase.
Using PRELOAD part would lead to problems when the problem
would happen before vg_write and vg_commit.
Also this change is necessary for snapshot creation sequence.
Failure to do so results in "Performing unsafe table load while X device(s) are
known to be suspended" errors. While fixing the problem in this way works and
is consistent with the way the mirror segment type does it, it would be nice
to find a solution that uses the generic suspend/resume calls.
Also included in this check-in are additions to the test suite that perform
conversions on RAID LVs under a snapshot. These tests are disabled for the
time being due to a kernel bug that is yet to be tracked down.
Similar to the "mirror" segment type's log device, _add_dev_to_dtree should
be called and not _add_lv_to_dtree when adding metadata sub-LVs to the deptree.
Since _add_lv_to_dtree was being called, 'origin_only' could be set if a
snapshot sits on top of the RAID device. This would cause the actual device
that needed to be added to be skipped in favor of the non-existant device,
"<foo>-real".
Reformat name and path how the LV is represented with lvm1 compatible option,
to switch to the old way - which had number of problem - i.e. many links
do not exist - since for private devices we are not creating them.
Add more info about thin pools and volumes.
Since striped name function knows when to report 'linear' instead of
'stripe' type name - drop it from this place.
This fixes problem when reporting segtype e.g. for thin-pool which
is also using area_count=1 to store thin data device reference.
It also returns properly strduped memory instead of badly casted const char*.
This patch to the suspend code - like the similar change for resume -
queries the lock mode of a cluster volume and records whether it is active
exclusively. This is necessary for suspend due to the possibility of
preloading targets. Failure to check to exclusivity causes the cluster target
of an exclusively activated mirror to be used when converting - rather than
the single machine target.
Since snapshot needs to suspend origin - it might lead to pool userspace
deadlock (as the pool will wait for new space in case it would be overfilled,
but dmeventd would not be able to resize it, as the lvcreate operation would
have kept the VG lock.)
To minimize the risk of such scenario - we prevent to create new snapshot
in case we are over the threshold - but beware, there is still small timewindow,
so keep threshold at some reasonable level!
New field Data% is able to display info about
thin_pool, thin, snapshot and has generic meaning here.
Simple Time/Host field are here to display host and time creation.
Basic support to keep info when the LV was created.
Host and time is stored into LV mda section.
FIXME: Current version doesn't support configurable string via lvm.conf
and used fixed version strftime "%Y-%m-%d %T %z".
This value returns percentage of 'mapped' size compared with total LV size.
(Without passed seg pointer it return highest mapped size - but it's
not used yet.)
used to be a few mis-ordered memory accesses (release and access in the next
block). Fix that set_flag could have sometimes corrupted the flags being
modified.
A few issues with metadata tracking are sorted out as well now, and there are
only a few problems remaining before we can integrate lvmetad, mostly on the
client side:
- metadata areas need to be tracked in lvmetad (most likely to be addressed
through an extension of metadata, meaning no special support in lvmetad would
be needed)
- non-udev scanning code needs to be taught about telling lvmetad about device
disappearance (pvscan most importantly)
- this last item also needs to mesh with metadata inconsistencies and
suddenly-incomplete volume groups (aux disable_dev in tests); udev-based
scanning should address this separately and more elegantly
Add 'blkdevname' and 'blkdevs_used' field to dmsetup info -c -o.
Add 'blkdevname' option to dmsetup ls --tree to see block device names.
Add '-o options' to dmsetup deps and ls to select device name type on output.
This is accomplished by reading associated sysfs information. For a dm device,
this is /sys/dev/block/major:minor/dm/name (supported in kernel version >= 2.6.29,
for older kernels, the behaviour is the same as for non-dm devices).
For a non-dm device, this is a readlink on /sys/dev/block/major:minor, e.g.
/sys/dev/block/253:0 --> ../../devices/virtual/block/dm-0.
The last component of the path is a proper kernel name (block device name).
One can request to read only kernel names by setting the 'prefer_kernel_name'
argument if needed.
LVM- prefix.
Try harder not to leave stray empty devices around (locally or remotely) when
reverting changes after failures while there are inactive tables.
If we know major:minor number of device (which is known after resume) we will
try to use sysfs to set/get read ahead parameters of device.
This avoid potential problem of blocking commands like 'dmsetup info' awaiting
for device being usable for open/close - i.e. overfilled thin pool may block
such command.
We want to keep this logic -
when LV is extend - extend the LV by at least given amount,
when LV is reduced - reduce the LV by at most given amount.
So for this the rounding needs to be used.
Current logic which seems to satisfy give rule is to round up all
extent values for LV resize upward except for values with '-' sign
that are round downward.
This patch also fixes the problem when lvextend --use-polices tried
to extend LV the by i.e. 20% - but the resulting 20% were smaller
the extent size thus before this patch no extension happened.
For snapshot, prepare whole command in front into private buffer.
Add also some missing '\n' for syslog messages.
For raid and mirror only convert creation of command line string.
This should avoid any unbound growth of mempool for dm_split_names.
Fix some minor outstading issue from thin plugin introduction -
Call dmeventd_lvm2_exit() in failpath for registration.
Add some missing '\n' in syslog messages.
Since the !(dev->flags & DEV_REGULAR) code path just called
dev_name_confirmed() which has just called 'stat()' inside,
remove duplicate second stat() call here.
When both path have identical prefix i.e. /dev/disk/by-id
skip 2 x lstat() for /dev /dev/disk /dev/disk/by-id
and directly lstat() only different part of the path.
Reduces amount of lstat calls on system with lots of devices.
Add dm_get_status_thin_pool and dm_get_status_thin functions to
parse 'params' argument which is received via dm_get_next_target.
Returns filed structure allocated from given mempool.
The RAID plug-in for dmeventd now calls 'lvconvert --repair' to address failures
of devices in a RAID logical volume. The action taken can be either to "warn"
or "allocate" a new device from any spares that may be available in the
volume group. The action is designated by setting 'raid_fault_policy' in
lvm.conf - the default being "warn".
Also, don't allow a splitmirror operation on a RAID LV that is already tracking
a split, unless the operation is to stop the tracking and complete the split.
Example:
~> lvconvert --splitmirrors 1 --trackchanges vg/lv /dev/sdc1
# Now tracking changes - image can be merged back or split-off for good
~> lvconvert --splitmirrors 1 -n new_name vg/lv /dev/sdc1
# ^ Completes split ^
If a split is performed on a RAID that is tracking an already split image and
PVs are provided, we must ensure that
1) the already split LV is represented in the PVs
2) we are careful to split only the tracked image
RAID is not like traditional LVM mirroring. LVM mirroring required failed
devices to be removed or the logical volume would simply hang. RAID arrays can
keep on running with failed devices. In fact, for RAID types other than RAID1,
removing a device would mean substituting an error target or converting to a
lower level RAID (e.g. RAID6 -> RAID5, or RAID4/5 to RAID0). Therefore, rather
than removing a failed device unconditionally and potentially allocating a
replacement, RAID allows the user to "replace" a device with a new one. This
approach is a 1-step solution vs the current 2-step solution.
example> lvconvert --replace <dev_to_remove> vg/lv [possible_replacement_PVs]
'--replace' can be specified more than once.
example> lvconvert --replace /dev/sdb1 --replace /dev/sdc1 vg/lv
LVM metadata knows only of striped segments - not linear ones.
The activation code detects segments with a single stripe and switches
them to use the linear target.
If the new lvm.conf setting is set to 0 (e.g. in a test script), this
'optimisation' is turned off.
Cleanup generated files from coverage testing.
Do not skip standard .o compilation for lib/not and lib/harness.
Make a bit longer string in harness to fit new shell/ in.
Simplify /api makefile and use SUBDIRS target for test dir.
Properly cleanup Makefiles with distclean in /test.
Use symbolic links for shell scripts for non-srcdir compilation.
Use gcc warning options only for .c -> .o compilation
So it makes the output more clear.
Do not use INCLUDES and DEFS for .o -> .so.
Do not use CFLAGS for deps .d generation.
bitset_t.c:39: warning: 'last' may be used uninitialized in this function
Compiler is not smart enough to see the code path which avoid using
unitialized 'last'.
tests from unit-tests/*/*_t.c (now under test/unit). The valgrind/pool test is
missing, since it's not really a unit test and probably not too valuable
either. Available via "make unit" (and if --enable-testing was passed to
configure, also executed by make check).
Remove FIXMES - there should not be any pool free call since
the memory pool is from device manager, and pool is detroyed
after the operation, so doing extra free here would not help here.
However lv_has_target_type() is using cmd mempool so here the extra
call for dm_pool_free makes sence.
"result_independent_of_operands: ((dev->dev & 0xfff00UL) >> 8) ==
18446744073709551615UL /* -1 */ is always false regardless of the values
of its operands (logical operand of if)."
'dev->dev' is set in dev-cache.c _insert() and it's not expectable
st_rdev would have '-1'
This code has been introduced with drbd support commit and code never
worked - so eliminated.
Avoid creation of target type name when it's longer then
DM_MAX_TYPE_NAME (noticed by static analyzer where the
sp.target_type might be missing '\0' at the end.)
Before patch:
$> dmsetup create long
0 1000 looooooooooooooooooooooooooong
^D
device-mapper: reload ioctl failed: Invalid argument
After patch:
$> dmsetup create xxx
0 1000 looooooooooooooooooooooooooong
Target type name looooooooooooooooooooooooooong is too long.
Command failed
Use static buffer instead of stack allocated buffer.
This reduces stack size usage of lvm tool and the
change is very simple.
Since the whole library is not thread safe - it should not
add any new problems - and if there will be some conversion
it's easy to convert this to use some preallocated buffer.
For write we do not need to hold memory locked.
This relaxes many conditions and avoid problems when allocating
a lot of memory for writting metadata buffers.
(In case of huge MDA size this would lead to mismatch between
locked and unlocked memory region size).
Add also internal check we are not writing in critical section.
Removal of an inactive origin removes also all related snapshots.
When we now support 'old' external snapshots with thin volumes,
removal of pool will not only drop all thin volumes, but as
a consequence also all snapshots - which might be seen a bit
unexpected for the user - so add a query to confirm such action.
lvremove -f will skip the prompt.
Update region_size only for mirror and raid targets.
This fixes warning messages when vg is using small
extent size like 1KiB and no mirror/raid is created,
but the user still got the message:
$> vgcreate -s 1K vg <pvs>
$> lvcreate -L10K vg
Using reduced mirror region size of 4 sectors
Since we support snapshots of thin volumes, we could have more layers,
so we have to check whether tpool layer is going to be inserted.
As the _add_segment_to_dtree() is the only place that adds tpool
segment, we may just check pointer (no strcmp for layer).
Switch to use seg_is_ function instead of lv_is_.
When a PV label write is deferred to a vg_write call (as introduced by a patch
in 2.02.86), the PV is flagged with the internal UNLABELLED_PV flag. However,
when calling vg_archive before vg_write, we still have the PV labelled with the
UNLABELLED_PV flag which was not recognised as a proper flag while exporting
VG metadata:
# vgcreate vg /dev/sda
No physical volume label read from /dev/sda
Metadata inconsistency: Not all flags successfully exported.
Metadata inconsistency: Not all flags successfully exported.
Writing physical volume data to disk "/dev/sda"
Physical volume "/dev/sda" successfully created
Volume group "vg" successfully created
udev may also need to be disabled if you didn't build it statically too.
dmeventd.static could be fixed with some more work but I don't really see the
point: without dlopen() it's useless, and if you have dlopen(), why not support
normal shared libraries too?
Any documentation less-detailed than Documentation/device-mapper is
dangerous for the non-trivial ctr lines. And anyway, this should be in s4
not here.
Use standard manpage style.
Keep options and commands in alphabetic order.
Added at least a very simply info about some other targets.
TODO: documenting targest needs far more work...
Remove DM_THIN_ERROR_DEVICE_ID from API.
Remove API warning.
Drop code that was using DM_THIN_ERROR_DEVICE_ID (already commented)
Remove debug message which slipped in through some previous commit.
Add filter which tries to check if scanned device is part
of active multipath.
Firstly, only SCSI major number devices are handled in filter.
Then it checks if device has exactly one holder (in sysfs) and
if it is device-mapper device and DM-UUID is prefixed by "MPATH-".
If so, this device is filtered out.
The whole filter can be switched off by setting
mpath_component_detection in lvm.conf.
https://bugzilla.redhat.com/show_bug.cgi?id=597010
Signed-off-by: Milan Broz <mbroz@redhat.com>
If the extent_size is smaller then the chunk_size we may try
to find better aligment (wasting less space).
i.e. using 4KB extent_size and 64KB chunk size will
lead to creation of 64KB aligned thin volume.
Since we finaly recognize thin creation only after
_determine_snapshot_type() - move _read_activation_params()
after it - so we can support lvcreate -an thin snapshot.
Always make sure table gets reloaded.
For now activate and deactivate pool volume if it's not active.
FIXME: we could do this only if we are sure some thin volume is alive.
that we can also test clustered volume groups (vgcreate -c y) in the test
suite. Unfortunately we can't make this the testing default since cluster
mirrors require further infrastructure, and snapshots probably don't work at
all. I'll eventually add a few test cases that create clustered VGs
specifically.
Support allocation of metadata from the same PV, if the VG
is build only from one PV.
As thinp is not mirror - we do not require 2 PVs
for basic thin usage as user is losing only perfomance.
Since activation of pool is now independent on thin activation,
user may do whatever he needs - thought preferable thin should stay alive,
but it it will be found inactivate, update_pool will bring the pool up.
Replace detach_pool_messages with update_pool_lv.
Move creation code from to 'if' condition into 1.
Ensure creation has finished all previous message operations.
Let's put the overlay device over real thin pool device.
So we can get the proper locking on cluster.
Overwise the pool LV would be activate once implicitely
and in other case explicitely, confusing locking mechanism.
This patch make the activation of pool LV independent on
activation of thin LV since they will both implicitely use
real -thin pool device.
A little code shuffling and adding support for
DM_THIN_ERROR_DEVICE_ID which might be eventually be used
for activation of thin which is going to be deleted.
For now we do not need it lvm.
Normally, restart simply means "stop and start" for systemd. However, if
we're installing new versions of the dmeventd binary/libdevmapper, we need
to restart dmeventd. This fails if we have some devices monitored - we need
to call "dmeventd -R" instead.
The "ExecReload" did not work quite well in some old versions of systemd,
systemd assumed that only the configuration is reloaded on "ExecReload",
not the whole binary itself so it lost track of dmeventd daemon (it lost new
dmeventd PID). This is fixed and seems to be working fine now with recent
versions of dmeventd.
All thins are created with the next activation and VG is updated
without messages. Only some basic commands works.
(i.e. lvcreate -an -V10 -T mvg/pool)
There can be some combination to confuse this system.
This functionality for snapshots is going to be interesting.
Add a new node flag send_messages that is used to simplify
test when to call _node_send_messages().
Add call to _node_send_messages when pool is deeper in the tree.
If something fails during creation of thin LV remove such LV
and deactivate in case it's been already tried to activate
(i.e. thin kernel driver fails for some reason.)
To ensure we properly handle LV cluster locking - explicitely do
not allow to change the availability of the thin pool that is in use
for some thin LV.
As soon as the thin volume is created the only way to activate pool
is via implicit dependency.
Ignore thinpool open count for lv/vgchange operations.
Before adding a new virtual segment to LV, check first whether
the last segment isn't already of the same type. In this case
extend last segment instead of creating the new one.
Thin volumes should have always only 1 virtual segment, but it
helps also to virtual snapshot or error segtype..
There should be no need for retry for our internal devices - it would be hinding
our own bug in the tree processing.
Update error messages to show also also device name.
No WHATS_NEW - in release fix.
Git commit ID 0864378250 was meant to disallow
'mirrored' logs for cluster mirrors. However, when add_mirror_log is used
to create the log (as is now the case when using 'lvcreate' or converting only
the log) the check is bypassed.
This patch adds the check to add_mirror_log.
pvck prints 'extra' character from the label since there is no '\0'
after the struct label entry and just uint64_t follows directly.
So avoid it by limiting 8 chars to be printed.
https://www.redhat.com/archives/lvm-devel/2011-January/msg00109.html
Signed-off-by: Paul Bolle <pebolle tiscali nl>
grep need -F to check what we really want to test.
Add better test for existing device.
Currently this test DOES NOT work with real /dev handle via udev
since our tool does not see such device listet through udev.
FIXME: We might be able to see it at least through dmsetup table and
use for lvm.
Since thin is not able to use _lv_insert_empty_sublvs,
remove its appearence from this function.
Start to use extend_pool() function for desired functionality
and modify lv_extend() for this.
Code in _lv_insert_empty_sublvs was not able to provide proper
initialization order for thin pool LV.
New function extend_pool() first adds metadata segment to pool LV which
is still visible. Such LV is activate and cleared.
Then new meta LV is created and metadata segments are moved there.
Now the preallocated pool data segment is attached to the pool LV
and layer _tpool is created. Finaly segment is marked as thin_pool.
This function could be useful for other _manip source files.
Use dm_list manipulation function for provided functionality,
which make the code more readable and avoid touching list
internal details here.
This was the intended behaviour, as described in the lvchange man page, so you
have complete control through volume_list in lvm.conf, but the code seems to
have been treating -ae as local-only for a very long time.
When DEBUG_MEM is used, the memory is trashed with extra pattern before real
free() is called, and as this memory was marked as non accessible when used with
valgrind, make it again usable.
Couple FIXMEs put into the code for parts of the code which may be
improved later, since we might be able to add 'lazy' device creation later.
For now require exclusive activation.
Split syntax for thin-pool since it cannot be fully matched with snapshot.
So to avoid more confusion - take thin support into separate line.
Though still significant updates are needed for thin provisioning.
Certain errno codes could be expected in some situations thus
add experimental support for them.
When expected errno is set after ioctl error - function skips error
printing and exits succefully.
Currently only useful for thin pool messages.
monitoring state of the logical volumes they are currently acting on.
Until now, every time a logical volume has been changed by a dmeventd plugin,
this plugin would have called back to dmeventd through the external FIFO
mechanism. I am fairly sure this was superfluous, inefficient and possibly even
dangerous.
The '--merge' option to lvconvert works on snapshots and RAID1. The man
pages correctly reflect this, but the CLI help output still used the term,
'SnapshotLogicalVolume'.
lvm part of messaging.
Each message is now stored it's own thin pool section:
message1 {
create = lv
}
Messages are queued to thin pool dm target when this target
is going to be resumed or used through some dependency.
Currently 'delete' message are purely queued and processed
with next thin pool resume operation (i.e. create_thin).
WARNING - thin provisioning support is developmental code.
Version 2 of the userspace log protocol accepts return information during the
DM_ULOG_CTR exchange. The return information contains the name of the log
device that is being used (if there is one). The kernel can then register the
device via 'dm_get_device'. Amoung other things, this allows for userspace to
assemble a correct dependency tree of devices - critical for LVM handling of
suspend/resume calls.
Also, update dm-log-userspace.h to match the kernel header associated with
this protocol change. (Includes a version inc.)
The upstream kernel version that this file mirrors has changed, here is the
commit message:
commit 86a54a4802df10d23ccd655e2083e812fe990243
Author: Jonathan Brassow <jbrassow@redhat.com>
Date: Thu Jan 13 19:59:52 2011 +0000
dm log userspace: add version number to comms
This patch adds a 'version' field to the 'dm_ulog_request'
structure.
The 'version' field is taken from a portion of the unused
'padding' field in the 'dm_ulog_request' structure. This
was done to avoid changing the size of the structure and
possibly disrupting backwards compatibility.
The version number will help notify user-space daemons
when a change has been made to the kernel/userspace
log API.
Signed-off-by: Jonathan Brassow <jbrassow@redhat.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
When verify_udev_operations was disable, code for stacking fs operation for
lvm links was completely disable - but this code was also used for collecting
information, that a new node is being created.
Add a new flag which is set when a creation of lv symlinks is requested which
should restore old behaviour of lv_info function, that has called fs_sync()
before quere for open count on device.
Code here is using thread write protected variable without locking.
So add locking, for proper synchronization and a FIXME, since the
code needs closer look.
Barrier is supposed to be used in situation like this
and replace tricky mutex usage, where mutex has been unlocked
by a different thread than the locking thread.
Usage of thread unprotected init_test is not correct and needs probably lvm lock
since it part of lvm library. Current implementation may probably fail with
test mode and actually create something unexpectedly (and vice versa).
Since default thread stack size is around 8MB and clvmd creates for now thread
for message, clvmd may easily reach multi GB size of in-memory locked pages
(runs with mlockall()).
This patch significantly reduces memory usage to just tens of MB,
and now different reasons are the cause of server overloading.
Now we are running out of free file descriptors mostly.
Go with just 64KiB for stack.
Closer inspection should be made, whether we actually need to play with
settings at all.
Since default stack size is 8MB and gets mapped via page locking thus,
it seems there is no big help with preallocation of stack to some value.
Properly detect if the filters were refreshed properly.
(May needs few more fixes ??)
Filter refresh may fail because it may be out of free file descriptors
when clvmd gets overloaded.
Replace usleep with pthread condition to increase speed testing
(for simplicity just 1 condition for all locks).
Use thread mutex also for unlock resource (so it wakes up awaiting
threads)
Better check some error states and return error in fail case with
unlocked mutex.
Using log_warn to report missing symlinks as warning, since the command
itself returns as successful, we should not produce log_error().
log_warn is better fit here.
Cosmetic, since r is already 0 for the error path, no need to assign it there,
and r is assigned to 1 after switch command.
Also makes the code more readable.
Workaround for the current code with big FIXME,
since proper solution for pvmove needs to be developed.
Commiting this only for the purpose to get cluster testing covered.
Example:
~> lvconvert --type raid1 vg/mirror_lv
Steps to convert "mirror" to "raid1"
1) Allocate a RAID metadata LV for each mirror image from the same PVs
on which they are located.
2) Clear the metadata LVs. This involves writing LVM metadata, so we don't
change any aspects of the mirror LV before this so that the user can easily
remove LVs from the failed convert attempt while retaining the original
mirror.
3) Remove the mirror log, if it exists.
4) Add metadata LVs to mirror LV
5) Rename mirror sub-lvs (s/mimage/rimage/)
6) Change flags and segtype from mirror to raid1
Example:
~> lvconvert --type raid1 -m 1 vg/lv
The following steps are performed to convert linear to RAID1:
1) Allocate a metadata device from the same PV as the linear device
to provide the metadata/data LV pair required for all RAID components.
2) Allocate the required number of metadata/data LV pairs for the
remaining additional images.
3) Clear the metadata LVs. This performs a LVM metadata update.
4) Create the top-level RAID LV and add the component devices.
We want to make any failure easy to unwind. This is why we don't create the
top-level LV and add the components until the last step. Should anything
happen before that, the user could simply remove the unnecessary images. Also,
we want to ensure that the metadata LVs are cleared before forming the array to
prevent stale information from polluting the new array.
A new macro 'seg_is_linear' was added to allow us to distinguish linear LVs
from striped LVs.
This patch allows a mirror to be extended without an initial resync of the
extended portion. It compliments the existing '--nosync' option to lvcreate.
This action can be done implicitly if the mirror was created with the '--nosync'
option, or explicitly if the '--nosync' option is used when extending the device.
Here are the operational criteria:
1) A mirror created with '--nosync' should extend with 'nosync' implicitly
[EXAMPLE]# lvs vg; lvextend -L +5G vg/lv ; lvs vg
LV VG Attr LSize Pool Origin Snap% Move Log Copy% Convert
lv vg Mwi-a-m- 5.00g lv_mlog 100.00
Extending 2 mirror images.
Extending logical volume lv to 10.00 GiB
Logical volume lv successfully resized
LV VG Attr LSize Pool Origin Snap% Move Log Copy% Convert
lv vg Mwi-a-m- 10.00g lv_mlog 100.00
2) The 'M' attribute ('M' signifies a mirror created with '--nosync', while 'm'
signifies a mirror created w/o '--nosync') must be preserved when extending a
mirror created with '--nosync'. See #1 for example of 'M' attribute.
3) A mirror created without '--nosync' should extend with 'nosync' only when
'--nosync' is explicitly used when extending.
[EXAMPLE]# lvs vg; lvextend -L +5G vg/lv; lvs vg
LV VG Attr LSize Pool Origin Snap% Move Log Copy% Convert
lv vg mwi-a-m- 20.00m lv_mlog 100.00
Extending 2 mirror images.
Extending logical volume lv to 5.02 GiB
Logical volume lv successfully resized
LV VG Attr LSize Pool Origin Snap% Move Log Copy% Convert
lv vg mwi-a-m- 5.02g lv_mlog 0.39
vs.
[EXAMPLE]# lvs vg; lvextend -L +5G vg/lv --nosync; lvs vg
LV VG Attr LSize Pool Origin Snap% Move Log Copy% Convert
lv vg mwi-a-m- 20.00m lv_mlog 100.00
Extending 2 mirror images.
Extending logical volume lv to 5.02 GiB
Logical volume lv successfully resized
LV VG Attr LSize Pool Origin Snap% Move Log Copy% Convert
lv vg Mwi-a-m- 5.02g lv_mlog 100.00
4) The 'm' attribute must change to 'M' when extending a mirror created without
'--nosync' is extended with the '--nosync' option. (See #3 examples above.)
5) An inactive mirror's sync percent cannot be determined definitively, so it
must not be allowed to skip resync. Instead, the extend should ask the user if
they want to extend while performing a resync.
[EXAMPLE]# lvchange -an vg/lv
[EXAMPLE]# lvextend -L +5G vg/lv
Extending 2 mirror images.
Extending logical volume lv to 10.00 GiB
vg/lv is not active. Unable to get sync percent.
Do full resync of extended portion of vg/lv? [y/n]: y
Logical volume lv successfully resized
6) A mirror that is performing recovery (as opposed to an initial sync) - like
after a failure - is not allowed to extend with either an implicit or
explicit nosync option. [You can simulate this with a 'corelog' mirror because
when it is reactivated, it must be recovered every time.]
[EXAMPLE]# lvcreate -m1 -L 5G -n lv vg --nosync --corelog
WARNING: New mirror won't be synchronised. Don't read what you didn't write!
Logical volume "lv" created
[EXAMPLE]# lvs vg
LV VG Attr LSize Pool Origin Snap% Move Log Copy% Convert
lv vg Mwi-a-m- 5.00g 100.00
[EXAMPLE]# lvchange -an vg/lv; lvchange -ay vg/lv; lvs vg
LV VG Attr LSize Pool Origin Snap% Move Log Copy% Convert
lv vg Mwi-a-m- 5.00g 0.08
[EXAMPLE]# lvextend -L +5G vg/lv
Extending 2 mirror images.
Extending logical volume lv to 10.00 GiB
vg/lv cannot be extended while it is recovering.
7) If 'no' is selected in #5 or if the condition in #6 is hit, it should not
result in the mirror being resized or the 'm/M' attribute being changed.
NOTE: A mirror created with '--nosync' behaves differently than one created
without it when performing an extension. The former cannot be extended when
the mirror is recovering (unless in-active), while the latter can. This is
a reasonable thing to do since recovery of a mirror doesn't take long (at
least in the case of an on-disk log) and it would cause far more time in
degraded mode if the extension w/o '--nosync' was allowed. It might be
reasonable to add the ability to force the operation in the future. This
should /not/ force a nosync extension, but rather force a sync'ed extension.
IOW, the user would be saying, "Yes, yes... I know recovery won't take long
and that I'll be adding significantly to the time spent in degraded mode, but
I need the extra space right now!".
This patch also does some clean-up of the splitmirrors code.
I've attempted to clean-up the splitmirrors code to make it easier to
understand with fewer operations. I've tried to reduce the number of
metadata operations without compromising the intermediate stages which
are necessary for easy clean-up in the even of failure.
These changes now correctly handle cluster situations - including exclusive
cluster mirrors. Whereas before, a splitmirror operation would result in
remote nodes having LVM commands report the newly split LV with a proper
name while DM commands would report the old (pre-split) names of the device.
IOW, there was a kernel/userspace mismatch.
The original commit comments can be located via this git commit ID:
7d8e615c0b
There were three possible solutions to the original problem proposed in the
initial check-in. The one chosen was as follows:
2) Do like _remove_mirror_images does and suspend the original, then suspend
the sub-lv (the error target), then resume the sub-lv, and finally resume the
original LV. This seems like extra pointless operations to me, but it doesn't
produce the error message (although, I'm not sure why) and it allows us to
leave the visible flag in place.
Turns out, the cluster also views the extra suspend/resume operations as
pointless too and ignores them. So, this solution doesn't work in a cluster.
Further, I've noticed that in addition to the remote cluster nodes still getting
I/O errors from scanning the error target, they also have a different LVM and
DM views of the same LV. IOW, while the LVM level (gotten from the LVM metadata)
sees the correct name for the newly split LV, device-mapper still maintains the
old names.
Because the original fix failed to completely fix the problem (or work-around it)
and because a better solution must be found to address the additional cluster
issue of device renaming, I am reverting the above mentioned commit.
The current code does not always assign proper udev flags to sub-LVs (e.g.
mirror images and log LVs). This shows up especially during a splitmirror
operation in which an image is split off from a mirror to form a new LV.
A mirror with a disk log is actually composed of 4 different LVs: the 2
mirror images, the log, and the top-level LV that "glues" them all together.
When a 2-way mirror is split into two linear LVs, two of those LVs must be
removed. The segments of the image which is not split off to form the new
LV are transferred to the top-level LV. This is done so that the original
LV can maintain its major/minor, UUID, and name. The sub-lv from which the
segments were transferred gets an error segment as a transitory process
before it is eventually removed. (Note that if the error target was not put
in place, a resume_lv would result in two LVs pointing to the same segment!
If the machine crashes before the eventual removal of the sub-LV, the result
would be a residual LV with the same mapping as the original (now linear) LV.)
So, the two LVs that need to be removed are now the log device and the sub-LV
with the error segment. If udev_flags are not properly set, a resume will
cause the error LV to come up and be scanned by udev. This causes I/O errors.
Additionally, when udev scans sub-LVs (or former sub-LVs), it can cause races
when we are trying to remove those LVs. This is especially bad during failure
conditions.
When the mirror is suspended, the top-level along with its sub-LVs are
suspended. The changes (now 2 linear devices and the yet-to-be-removed log
and error LV) are committed. When the resume takes place on the original
LV, there are no longer links to the other sub-lvs through the LVM metadata.
The links are implicitly handled by querying the kernel for a list of
dependencies. This is done in the '_add_dev' function (which is recursively
called for each dependency found) - called through the following chain:
_add_dev
dm_tree_add_dev_with_udev_flags
<*** DM / LVM divide ***>
_add_dev_to_dtree
_add_lv_to_dtree
_create_partial_dtree
_tree_action
dev_manager_activate
_lv_activate_lv
_lv_resume
lv_resume_if_active
When udev flags are calculated by '_get_udev_flags', it is done by referencing
the 'logical_volume' structure. Those flags are then passed down into
'dm_tree_add_dev_with_udev_flags', which in turn passes them to '_add_dev'.
Unfortunately, when '_add_dev' is finding the dependencies, it has no way to
calculate their proper udev_flags. This is because it is below the DM/LVM
divide - it doesn't have access to the logical_volume structure. In fact,
'_add_dev' simply reuses the udev_flags given for the initial device! This
virtually guarentees the udev_flags are wrong for all the dependencies unless
they are reset by some other mechanism. The current code provides no such
mechanism. Even if '_add_new_lv_to_dtree' were called on the sub-devices -
which it isn't - entries already in the tree are simply passed over, failing
to reset any udev_flags. The solution must retain its implicit nature of
discovering dependencies and be able to go back over the dependencies found
to properly set the udev_flags.
My solution simply calls a new function before leaving '_add_new_lv_to_dtree'
that iterates over the dtree nodes to properly reset the udev_flags of any
children. It is important that this function occur after the '_add_dev' has
done its job of querying the kernel for a list of dependencies. It is this
list of children that we use to look up their respective LVs and properly
calculate the udev_flags.
This solution has worked for single machine, cluster, and cluster w/ exclusive
activation.
The problem as reported by "ben <benscott@nwlink.com>" on lvm-devel:
vgsplit fails with mirrored mirror log
#lvs --all -o lv_name,lv_attr,devices
LV Attr Devices
MyMirror mwi--
[MyMirror_mimage_0] Iwi--- /dev/sdq(0)
[MyMirror_mimage_1] Iwi--- /dev/sdo(0)
[MyMirror_mimage_2] Iwi--- /dev/sdi(0)
[MyMirror_mlog] mwi---
[MyMirror_mlog_mimage_0] Iwi--- /dev/sds(0)
[MyMirror_mlog_mimage_1] Iwi--- /dev/sde(0)
#vgsplit -v "TestA" "TestB" "/dev/sdq" "/dev/sdo" "/dev/sdi" "/dev/sds"
"/dev/sde"
Checking for volume group "TestA"
Checking for new volume group "TestB"
Archiving volume group "TestA" metadata (seqno 213).
Can't split mirror MyMirror between two Volume Groups
AFTER FIX:
[root@bp-01 ~]# lvs -a -o name,vg_name,devices vg new
Volume group "new" not found
Skipping volume group new
LV VG Devices
lv vg lv_mimage_0(0),lv_mimage_1(0)
[lv_mimage_0] vg /dev/sdb1(0)
[lv_mimage_1] vg /dev/sdc1(0)
[lv_mlog] vg lv_mlog_mimage_0(0),lv_mlog_mimage_1(0)
[lv_mlog_mimage_0] vg /dev/sdh1(0)
[lv_mlog_mimage_1] vg /dev/sdi1(0)
[root@bp-01 ~]# vgsplit vg new /dev/sd[bchi]1
New volume group "new" successfully split from "vg"
[root@bp-01 ~]# lvs -a -o name,vg_name,devices vg new
LV VG Devices
lv new lv_mimage_0(0),lv_mimage_1(0)
[lv_mimage_0] new /dev/sdb1(0)
[lv_mimage_1] new /dev/sdc1(0)
[lv_mlog] new lv_mlog_mimage_0(0),lv_mlog_mimage_1(0)
[lv_mlog_mimage_0] new /dev/sdh1(0)
[lv_mlog_mimage_1] new /dev/sdi1(0)
Make limits for thin data_block_size and device_id part of public API.
FIXME: read them possible from some kernel header file in the future ?
But we may need to support different values for different versions ?
Since execve passed only NULL as environ, we had lost all environment vars on
restart - thus actually running 'different' clvmd then the one at start.
Preserving environ allows to restart clvmd with the same settings
(i.e. LD_LIBRARY_PATH)
Add test for second restart.
Since it's internal function and we always check for NULL value
before call - this is safe.
Just for case add nonnull attribute so analyzer might better
catch error.
It's 100% equivalent test - since it always happen for the first iteration.
But the check for 'l' is understandable with analyzers - since analyzer
is not smart enough to deduce connection between root->child == NULL.
Add named cluster_ops to easily learn the name of the active cluster manager,
so we are able to restart singlenode manager in testing.
Add simple test for clvmd -S (restart) and -R (refresh)
(though it needs some extensions).
When running tests it might be useful to have an override option when
testing on real /dev and some broken system (i.e. Debian and its rules).
So one can use:
LVM_TEST_DEVDIR=/dev LVM_VERIFY_UDEV=1 make check
When read in drain returned <0 value, terminal content has been trashed.
Remove unneeded memset() and use whole buffer.
Free readbuf before exit (valgrind).
Next iteration for better fit of lvmetad compilation.
Move build of libdaemon.a into common subdir Makefile.
libdaemon.a is device-mapper target.
Build and install lvmetad as lvm2 target.
Test whether nodes could be used on given filesystem where TMP
dir is being used and skip teardown quicker in fail case.
(makes the problem quickly obvious if you try to such fs).
Skip teardown_dev if we have not created any devs yet.
and do not mkdir /dev/mapper dir when LVM_TEST_DEVDIR is set.
Drop this test from t-000-basic.sh.
Read 2 environmental vars to learn about overide position for
CLVMD and LVM binaries.
We support LVM_BINARY in other script - and this way we could easily
test restart in our test-suite.
Bugfix:
Add (most probably unfinished) support for -E arg with list of exclusive
locks. (During clvmd restart all exclusive locks would have been lost and
in fact, if there would have been an exclusive lock, usage text would be
printed and clvmd exits.)
Instead of parsing list options multiple times every time some lock UUID is
checked - put them straight into the hash table - make the code easier to
understand as well.
Remove was_ex_lock() function (replaced with dm_hash_lookup()).
Swap return value for get_initial_state() (1 means success).
Update man pages and usage info for -E option.
Before, we used to display "Can't remove open logical volume" which was
generic. There 3 possibilities of how a device could be opened:
- used by another device
- having a filesystem on that device which is mounted
- opened directly by an application
With the help of sysfs info, we can distinguish the first two situations.
The third one will be subject to "remove retry" logic - if it's opened
quickly (e.g. a parallel scan from within a udev rule run), this will
finish quickly and we can remove it once it has finished. If it's a
legitimate application that keeps the device opened, we'll do our best
to remove the device, but we will fail finally after a few retries.
Add dm_device_has_mounted_fs fn to check mounted filesystem on a device.
This requires sysfs directory to be correctly set via dm_set_sysfs_dir
(/sys by default). If sysfs dir is not used or it's set incorrectly,
dm_device_has_{holders,mounted_fs} will return 0!
If you specify the segment type (e.g. --type mirror) and the mirrors argument
as zero, it would result in a mirrored LV with only one image. While the device
may be valid in theory, it should not be allowed in practice. It also makes it
difficult on the conversion tools, since they react badly to single-image
mirrors.
seg->areas and seg->meta_areas. We also need to copy the memory from the
old arrays to the newly allocated arrays. The amount of memory to copy was
determined by seg->area_count. However, seg->area_count was being set to the
higher value after copying the 'seg->areas' information, but before copying
the 'seg->meta_areas' information. This means we were copying more memory
than necessary for 'seg->meta_areas' - something that could lead to a segfault.
Patch fixes Clang warnings about possible access via lv_name NULL pointer.
Replaces allocation of memory (strdup) with just pointer assignment
(since execve is being called anyway).
Checks for !*lv_name only when lv_name is defined.
(and as I'm not quite sure what state this really is - putting a FIXME
around - as this rather looks suspicios ??).
Add debug print of passed clvmd args.
(since data in statbuf are invalid).
Check whether sysconf managed to find _SC_PAGESIZE.
Report at least debug warning about failing unlink
(logging scheme here seems to be a different then in lvm).
Duplicate terminal FDs and use similar code as is made in clvmd
and cleanup warns about missing open/close tests.
FIXME: Looks like we already have 3 instancies of the same code in lvm repo.
It seems like test-suite is significantly slowed down on removal phase.
Maybe it should be part of generic teardown - though the reason of slowdown
should be also discovered - probably related to retry loop ?
When the srcdir == builddir we get the link on non-exectable file.
So make always sure fsadm.sh is executable script file.
Add cleanup target for lib/fsadm.
Compiler says variable may be used uninitialized. It can't be, but we
initialize the variable to NULL anyway. Also, remove the double initialization
of another variable.
When fsadm is test - it needs to execute lvm and fsadm from non-standard path
setting. So adding a support in fsadm script when user set LVM_BINARY, then
the lvm command invoced from fsadm will have the same PATH setting as before
entering fsadm command.
Needed for testing.
In case someone would use filename paths with spaces when changing
this script surround commands with '"'.
With default settings there is no change in behavior.
This bug showed up when trying to add a log to a mirror whose images are on
multiple devices. This is an intra-release regression and no WHATS_NEW
entry will be added. The error was introduce in the following commit:
2d8a2f35c7
The solution is to recognise in _alloc_init that if there are no mirrors
or stripes specified, then 'new_extents' should be zero.
to settle udev before calling deactivate_lv.
This is an intra-release regression (no WHATS_NEW entry required). It is
part of the fix for the current WHATS_NEW entry:
Work around resume_lv causing error LV scanning during splitmirror operation.
Currently test is skipped by default (since it needs code hack to work)
Check command line options to create & remove thin pools and thin volumes.
Activation code for thin LV support is missing, thus it only works without
driver loaded.
When user wants to remove thin pool - check if there are no thin volumes using it.
If so - query before removal (or -ff for no question) and remove them first.
LVM has huge set of options now - it's approaching 60 short-arg less options
and we get interesting case of misdetection for 'merge' option which has been
put into the middle of options with 'short_arg' - thus certainly past 65. (ASCII 'A').
To avoid confusion of short_arg with long_opt number - add '128' to all such
non-short-arg options.
When LV is unlinked, we want to catch problem in vg_validate,
that LV has changed.
i.e. catch LV has been removed and is no long thin_pool while still
being referenced by some thin volume.
Transfer of build_dm_uuid() function into libdm made uuid_prefix as parameter,
thus sizeof() was replaced with strlen() and room for '\0' missed.
As it's only fix in current version - no whatsnew.
Revert John patch, which fixed only 1 place where ~LVM_WRITE was in use and
convert ommited LVM_READ/WRITE flags to 64bit constants as well.
(Since both 'status' flags for LV and VG are 64bit.)
Changing lv_mirror_count to only count the AREA_LVs made the function
stop working for PVMOVE mirrors. A conditional has been added to fix
that problem. Additionally, when counting the images in a mirror stack,
we don't need to subtract 1 from the count we get back from the
lv_mirror_count call on the temporary mirror layer. (This is because we
are no falsely counting the top layer of the temporary mirror.)
lv_mirror_count was not able to handle mirrors of stripes properly. When a
failed device is removed, the MIRRORED status flag is removed from the LV
conditionally based on the results of lv_mirror_count. However, lv_mirror_count
trusted the MIRRORED flag - thinking any such LV must be mirrored. It would
happily assign first_seg(lv)->area_count as the number of mirrors, but when
a mirrored striped LV was reduced to a simple striped LV area_count would be
the number of /stripes/ not the number of /mirrors/. A result higher than 1
would be returned from lv_mirror_count, the MIRRORED flag would not be cleared,
and the LV would fail to be up-converted properly in lvconvert_mirrors_aux
because of it.
The operation of deactivating the residual error target LV after removing a
mirror layer can cause a "device in-use" conflict with udev. Giving udev a
poke before calling deactivate_lv eliminates the conflict. The stick used
to poke udev is 'sync_local_dev_names'.
Kernel requires a mirror to be at least 1 region large. So,
if our mirror log is itself a mirror, it must be at least
1 region large. This restriction may not be necessary for
non-mirrored logs, but we apply the rule anyway.
(The other option is to make the region size of the log
mirror smaller than the mirror it is acting as a log for,
but that really complicates things. It's much easier to
keep the region_size the same for both.)
WHATS_NEW entry:
Fix log size calculation when only a log is being added to a mirror.
The original fix pass the mirror LV to allocate_extents (rather than
passing NULL) so that _alloc_init could correctly determine the necessary
size of the mirror log. In the previous check-in, I noted:
In order to get a decent value computed, we need to pass in the 'lv' argument
to allocate_extents. This would normally imply a desire for cling/contiguous
allocation to the given LV, but since we are not allocating any parallel
extents and only log extents, it works fine.
However, passing in the LV did have unintended consequences on the placement of
the log. The better solution is to pass in the number of extext that are in
the mirror LV instead of the LV itself. This will not cause the allocator to
reserve that number of extents, because 'stripes' and 'mirrors' are specified
as 0. Thus, 'extents' is used to calculate the size of the log, but won't
affect how much is allocated.
LVM_WRITE is a 32-bit flag. Now that RAID[_IMAGE|_META] are 64-bit,
and'ing a RAID LV's status against LVM_WRITE can reset the higher order
flags.
A similar thing will affect thinp flags if not careful.
This is a workaround for long-lasting problem with using the WATCH udev
rule. When trying to remove a DM device, this one can still be opened
while processing the event in parallel (generated based on the WATCH
udev rule).
Let's use this until we have a proper solution.
_alloc_init calculates the number of necessary log extents via
'mirror_log_extents'. 'mirror_log_extents' takes 3 arguments: region_size,
pe_size, and size of the mirror LV. Unfortunately, _alloc_init is guessing at
the mirror size by using 'ah->new_extents / ah->area_multiple' - the number of
extents that the mirror images have. However, this is /always/ wrong when
allocating the log separately. Further, the log is always allocated separately
unless we are up-converting the mirror at the same time. It was by luck alone
that a default value of '1' reflects what we want in most cases.
In order to get a decent value computed, we need to pass in the 'lv' argument
to allocate_extents. This would normally imply a desire for cling/contiguous
allocation to the given LV, but since we are not allocating any parallel
extents and only log extents, it works fine.
When an image is split from a 2-way mirror, the original mirror is converted to
a linear device. To do this, the top "layer" must be removed. The segments
are transferred from the sub-lv to the top-level LV and the link is severed.
The former sub-lv - having its segments transferred - now contains a temporary
error target.
When the original LV is resumed, the old sub-lv that now contains an error
segment is activated and scanned. This is what causes the I/O error messages.
There are three ways to fix this problem:
1) Do not set the sub-lv which contains the error target as "visible" before
suspending the original LV. This way, when the original is resumed, the sub-lv
device node is not created and it is not scanned - avoiding the error messages.
The problem with this approach is that if the machine crashes after the
resume, it leaves the *hidden* LV in place and the user has a more difficult
time noticing that it needs to be cleaned up. Thus, this type of processing is
frowned upon.
2) Do like _remove_mirror_images does and suspend the original, then suspend
the sub-lv (the error target), then resume the sub-lv, and finally resume the
original LV. This seems like extra pointless operations to me, but it does not
produce the error message (although, I'm not sure why) and it allows us to
leave the visible flag in place.
3) Flag the sub-lv (error target) with a "do not scan" flag. This seems like
the cleanest approach, but I have been unable to find the method for doing
this. LVs get tagged in such a way by _get_udev_flags, but in this case the
resume of the original LV also resumes the error target LV without running it
through _get_udev_flags (likely because they are no longer linked). Could
there be something wrong in resume_lv?
Option #2 was chosen to fix this bug, but it seems like more of a workaround
for now.
A gentle reminder that anyone relying on the output of reporting commands
like lvs in scripts must use -o to guarantee they get the fields they expect.
The default sequence of fields can change from release to release.
Equally, the 'attr' fields can have new values introduced and/or characters
appended to them.
Some major distributions are still using 'mawk' and they are not using
the latest version - we end here with hidden dependency on the latest
version of mawk (1.3.4) while i.e. Debian Lenny seems to stay with 1.3.3.
So we end with completely broken vgimportclone script on such system.
We would need to check for proper support of :space: and abort build if
it doesn't work or simplier replace [:space:] with [ \t] which seems
sufficient to make it work (as can be seen in this patch)
A better fix would be to use command line parameter override - leaving
as FIXME comment.
This patch makes t-vgimportclone.sh test passing on Lenny.
Makes dumpconfig whole-section output wrong in a different way from before,
but we should be able to merge cft_cmdline properly into cmd->cft now and
remove cascade.
There was a bad sequence:
*) Make changes to LV layout to split images (e.g. 4-way -> 2-way/2-way)
1) vg_write, suspend_lv(original_mirror), vg_commit
2) activate_lv(newly_split_lv)
3) resume_lv(original_mirror)
Step #2 is not allowed. However, without it, the resume of the original
mirror will also resume its former sub-LVs - making it impossible to
activate the newly split LV due to the changes in layering, pointers, and
names that had already been made. Additionally, the resume or the original
brings the sub-lv's online with names that differ from the metadata on disk -
also a no-no. Thus, the split must be done in stages such that the active LVs
always reflect what is in the committed LVM metadata.
First, alter the original mirror by releasing the images. The images are made
visible and independent as an intermediate stage. (This way, we can have
consistency between LVM metadata and active LVs.) The second stage collects
the recently split LVs, deactivates them, forms them into a mirror if necessary,
and then activates them. It is a bit of a circuitous method, but it is the only
way to split a mirror from a mirror and obey these general rules:
1) Never [de]activate sub-lvs when the top-level LV is suspended
2) Avoid having active LVs that differ from the description in the LVM metadata
Signed-off-by: Jonathan Brassow <jbrassow@redhat.com>
functionality. A number of bugs (copied and pasted all over the code) should
disappear:
- most string lookup based on dm_config_find_node would segfault when
encountering a non-zero integer (the intention there was to print an
error message instead)
- check for required sections in metadata would have been satisfied by
values as well (i.e. not sections)
- encountering a section in place of expected flag value would have
segfaulted (due to assumed but unchecked cn->v != NULL)
daemon/common code in a single libdaemon.a, which is completely private. This
is currently linked into the lvmetad binary, and will be linked into LVM (the
client part, since static linking only picks up only symbols that are actually
used). I have also added --enable/disable-lvmetad to ./configure; although the
current default is off, I expect this to be flipped to on shortly. There's no
LVM-side support yet, but when there is, even when built, it'll still need to
be enabled by an lvm.conf option.
leaving behind the LVM-specific parts of the code (convenience wrappers that
handle `struct device` and `struct cmd_context`, basically). A number of
functions have been renamed (in addition to getting a dm_ prefix) -- namely,
all of the config interface now has a dm_config_ prefix.
There's a very high memory usage when calling _pv_analyse_mda_raw (e.g. while
executing pvck) that can end up with "out of memory".
_pv_analyse_mda_raw scans for metadata in the MDA, iteratively increasing the
size to scan with SECTOR_SIZE until we find a probable config section or we're
at the edge of the metadata area. However, when using a memory pool, we're also
iteratively chasing for bigger and bigger mempool chunk which can't be found
and so we're always allocating a new one, consuming more and more memory...
This patch just changes the mempool to direct memory allocation in this
problematic part of the code.
There is duplicate code in t-lvconvert-raid.sh and t-lvcreate-raid.sh.
This should be moved into a common file which is then sourced by these two
files. I'll wait to move the duplicate code until I can talk to mornfall.
This patch adds the ability to upconvert a raid1 array - say from 2-way to
3-way. It does not yet support upconverting linear to n-way.
The 'raid' device-mapper target allows for individual components (images) of
an array to be specified for rebuild. This mechanism is used when adding
new images to the array so that the new images can be resync'ed while the
rest of the images in the array can remain 'in-sync'. (There is no
mirror-on-mirror layering required.)
~> lvconvert --splitmirrors 1 --trackchanges vg/lv
The '--trackchanges' option allows a user the ability to use an image of
a RAID1 array for the purposes of temporary read-only access. The image
can be merged back into the array at a later time and only the blocks that
have changed in the array since the split will be resync'ed. This
operation can be thought of as a partial split. The image is never completely
extracted from the array, in that the array reserves the position the device
occupied and tracks the differences between the array and the split image via
a bitmap. The image itself is rendered read-only and the name (<LV>_rimage_*)
cannot be changed. The user can complete the split (permanently splitting the
image from the array) by re-issuing the 'lvconvert' command without the
'--trackchanges' argument and specifying the '--name' argument.
~> lvconvert --splitmirrors 1 --name my_split vg/lv
Merging the tracked image back into the array is done with the '--merge'
option (included in a follow-on patch).
~> lvconvert --merge vg/lv_rimage_<n>
The internal mechanics of this are relatively simple. The 'raid' device-
mapper target allows for the specification of an empty slot in an array
via '- -'. This is what will be used if a partial activation of an array
is ever required. (It would also be possible to use 'error' targets in
place of the '- -'.) If a RAID image is found to be both read-only and
visible, then it is considered separate from the array and '- -' is used
to hold it's position in the array. So, all that needs to be done to
temporarily split an image from the array /and/ cause the kernel target's
bitmap to track (aka "mark") changes made is to make the specified image
visible and read-only. To merge the device back into the array, the image
needs to be returned to the read/write state of the top-level LV and made
invisible.
Users already have the ability to split an image from an LV of "mirror"
segtype. This patch extends that ability to LVs of "raid1" segtype.
This patch only allows a single image to be split off, however. (The
"mirror" segtype allows an arbitrary number of images to be split off.
e.g. 4-way => 3-way/linear, 2-way/2-way, linear,3-way)
of top-level LV.
We can't activate sub-lv's that are being removed from a RAID1 LV while it
is suspended. However, this is what was being used to have them show-up
so we could remove them. 'sync_local_dev_names' is a sufficient and
proper replacement and can be done after the top-level LV is resumed.
1) add new function 'raid_remove_top_layer' which will be useful
to other conversion functions later (also cleans up code)
2) Add error messages if raid_[extract|add]_images fails
3) Add function prototypes to prevent compiler warnings when
compiling with '--with-raid=shared'
Fix a couple more issues that kabi found.
- Add some error messages in failure cases
- s/malloc/zalloc/
- use vg->vgmem for lv names instead of vg->cmd->mem
Skip decoding of DM flags when device is removed.
We currently need DM flags only for add|change events. So forking
dmsetup process for removed devices is a waste of CPU time.
Udev is already quite slow, so make it just a tiny bit faster.
Add config option to enable crc checking of VG structures.
Currently it's disabled by default.
For the internal test-suite this check it is enabled.
Note: In the case the internal error is detected, debug build with
compile option DEBUG_ENFORCE_POOL_LOCKING helps to catch the source
of the problem.
Use debug pool locking functionality. So the command could check,
whether the memory in the pool has not been modified.
For lv_postoder() instead of unlocking and locking for every changed
struct status member do it once when entering and leaving function.
(mprotect would trap each such memory access).
Currently lv_postoder() does not modify other part of vg structure
then status flags of each LV with flags that are reverted back to
its original state after function exit.
Adding debuging functionality to lock and unlock memory pool.
2 ways to debug code:
crc - is default checksum/hash of the locked pool.
It gets slower when the pool is larger - so the check is only
made when VG is finaly released and it has been used more then
once.Thus the result is rather informative.
mprotect - quite fast all the time - but requires more memory and
currently it is using posix_memalign() - this could be
later modified to use dm_malloc() and align internally.
Tool segfaults when locked memory is modified and core
could be examined for faulty code section (backtrace).
Only fast memory pools could use mprotect for now -
so such debug builds cannot be combined with DEBUG_POOL.
Extend vginfo cache with cached VG structure. So if the same metadata
are use, skip mda decoding in the case, the same data are in use.
This helps for operations like activation of all LVs in one VG,
where same data were decoded giving the same output result.
Patch adds 1-to-1 connection between volume_group and lvmcache_vginfo.
Move the free_vg() to vg.c and replace free_vg with release_vg
and make the _free_vg internal.
Patch is needed for sharing VG in vginfo cache so the release_vg function name
is a better fit here.
As this flag could not have been set by the current code - removing it.
Note: because of the wrong code logic this call:
lvmcache_update_vg(correct_vg, correct_vg->status & PRECOMMITTED &
(inconsistent ? INCONSISTENT_VG : 0));
had always passed '0' - now after flag removal it's passing
PRECOMMITTED flag in - this present functinal change in this patch.
To match the original functionality - 0 had to be always passed.
More testing is needed here.
Compiler complaining that meta_lv could be used uninitialized. (Not true
because it is protected by 'clear_metadata'.) I switched to using 'lv->vg',
as it makes no difference to vg_[write|commit].
(here clvmd crashed in the middle of operation),
lock is not removed from cache - here is one example:
locking/cluster_locking.c:497 Locking VG V_vg_test UN (VG) (0x6)
locking/cluster_locking.c:113 Error writing data to clvmd: Broken pipe
locking/locking.c:399 <backtrace>
locking/locking.c:461 <backtrace>
Internal error: Volume Group vg_test was not unlocked
Code should always remove lock info from lvmcache and update counters
on unlock, even if unlock fails.
Today, we use "suppress_messages" flag (set internally in init_locking fn based
on 'ignorelockingfailure() && getenv("LVM_SUPPRESS_LOCKING_FAILURE_MESSAGES")'.
This way, we can suppress high level messages like "File-based locking
initialisation failed" or "Internal cluster locking initialisation failed".
However, each locking has its own sequence of initialization steps and these
could log some errors as well. It's quite misleading for the user to see such
errors and warnings if the "--sysinit" is used (and so the ignorelockingfailure
&& LVM_SUPPRESS_LOCKING_FAILURE_MESSAGES environment variable). Errors and
warnings from these intermediary steps should be suppressed as well if requested.
This patch propagates the "suppress_messages" flag deeper into locking init
functions. I've also added these flags for other locking types for consistency,
though it's not actually used for no_locking and readonly_locking.
Last usage was removed in Petr's commit related to VG mda repair fix
where relaxed check starts to ignore inconsistencies coming from
PVs that are marked MISSING - thus removing unused variable.
Defer the test of the function return value after the string memory is released.
Otherwise in this error path the string would present memory leak.
(Thought in this case we are already out of memory...)
Implementation described in doc/lvm2-raid.txt.
Basic support includes:
- ability to create RAID 1/4/5/6 arrays
- ability to delete RAID arrays
- ability to display RAID arrays
Notable missing features (not included in this patch):
- ability to clean-up/repair failures
- ability to convert RAID segment types
- ability to monitor RAID segment types
This should be set by default! Normally we have "activation/udev_sync = 1"
in lvm.conf (example.conf.in). But if we use lvm2 without any config file
(or without a definition within '--config' option) the DEFAULT_UDEV_SYNC
is used instead. Together with verify_udev_operations=0 (when we rely on
udev fully), this can cause races as the node could be missing when needed.
(See also https://bugzilla.redhat.com/show_bug.cgi?id=723144)
Systemd preloads file descriptors for us and passes them in for
newly spawned daemon when using on-demand fifo (or socket)
based activation.
This patch adds checks for file descriptors preloaded by
systemd and uses them instead of opening the FIFOs again
to properly support on-demand FIFO-based activation.
(We'll change FIFOs to sockets soon - but still this
part of the code will stay almost the same.)
The filename to adjust the oom score was changed in 2.6.36.
We should use oom_score_adj instead of oom_adj (which is still
there under /proc, but it's scheduled for removal in August 2012).
New oom_score_adj uses a range from -1000 (OOM_SCORE_ADJ_MIN,
disable oom killing) to 1000 (OOM_SCORE_ADJ_MAX).
Clvmd detects modifed config file before it takes lv_lock.
If the config file is changed rapidly - the change was ignored within
a seocnd ranged. This patch adds also compare of file size.
So change like some flag for 0 to 1 would pass unnoticed - but
it's quick fix for failing test suite.
FIXME: Implement inotify solution.
very reasonable amount of parallel access, although the hash tables may become
a point of contention under heavy loads. Nevertheless, there should be orders
of magnitude less contention on the hash table locks than we currently have on
block device scanning.
where the naming is left completely on lvm.
(Commited code has been different version of test).
So here it should be able to figure out new free name and create a new LV.
device that has been brought back from the dead: this sometimes fails with
clvmd (the cache is updated "too soon"). Instead, force a pvscan and rely on an
up-to-date cache as usual.
If MD linear device has set rounding (overload chunk size attribute),
the pvcreate command prints this warning:
/dev/md0 sysfs attr level not in expected format: linear
teardown sequence. (Previously the snapshot was deactivated while its
origin was active and before its removal was committed to disk, so
restarting after a crash at the point would leave corruption.)
When some target is passing empty parameters to some dm target,
report this as an internal error to better catch some broken
table construction (some mirror conversions seem to be doing
this for now).
Since some test may leave devices in suspend mode which would require
carefull order of resume operation - use '-f' to replace them with
error targets
For disable_dev - when 'error' target is used for open count - treat
return code as ok (|| true) to avoid breaking futher test processing.
The conditional is not just unnecessary, it would have been wrong. The code
is suppose to be checking if the 'splitmirrors_ARG' is negative, but it
instead is checking 'mirrors_ARG'. Rather than changing the argument being
checked, I've pulled the check entirely because 'splitmirrors_ARG' is already
guarenteed to not be negative by virtue of the fact that it is a 'int_arg'.
Negative values will be caught in _process_command_line().
When an LVM mirror is up-converted (an additional image added), it creates
a temporary mirror stack. The lower-level mirror in the stack that is
created was not being activated exclusively - violating the exclusive nature
of the original mirror. We now check for exclusive activation of a mirror
before converting it, and if found, we ensure that the temporary mirror
is also exclusively activated.
Mirrors used to be created by first creating a linear device and then adding
the other images plus the log. Now mirrors are created by creating all the
images in one go and then adding the log separately. The new way ran into
the condition that cluster mirrors cannot change the log type (in the case
of creation, from core -> disk) while the mirror is not active. (It isn't
active because it is in the process of being created.) The reason this
condition is in place is because a remote node may have the mirror active, and
we don't want to alter the log underneath it.
What we really needed was a way of checking if the mirror was active remotely
but not locally, and in that case do not allow a change of the log. I've added
this check, and cluster mirrors can now be created again.
This fn calls rm_dev_node directly - an exceptional case. It needs to check
the DM_UDEV_DISABLE_LIBRARY_FALLBACK flag directly (it's called in dm_task_run
normally where it's checked already).
We've used udev fallback code till now to check whether udev
created/removed the entries in /dev correctly and if not,
a repair was done (giving a warning messagea about that).
This patch adds a possibility to enable this additional check
and subsequent fallback only when required (debugging purposes
mostly) and trust udev completely.
So let's disable the fallback code by default and add a new
configuration option "activation/udev_fallback".
(The original code for creating the nodes will still be used
in case the device directory that is set in lvm.conf differs
from the one that udev uses and also when activation/udev_rules
is set to 0 - otherwise we would end up with no nodes/symlinks
at all)
It's useful to keep the partial flag cached - so just move the call
for vg_mark_partil_lvs() into import_vg_from_config_tree() so it gets
evaluated before it goes through the lvmcache.
This patch should not present any functional change.
Note: It is rather temporal solution - proper place is probably inside the
'read' call back - but needs some more discussion.
For now using this minor hack.
As the ACTIVATE_EXCL could be set only in clvmd code - there is no
use for this test in lv_add_mirrors() function only called from
tools context.
FIXME: Add cluster test case for this.
To avoid modification of 'read-only' volume group structure
add a new structure to pass local data around the code for LV
activation.
As origin_only is one such flag - replace this parameter with new
struct lv_activate_opts.
More parameters might eventually become part of lv_activate_opts.
transient error), stemming from the following sequence of events:
1) devices fail IO, triggering repair
2) dmeventd starts fixing up the mirror
3) during the downconversion, a new metadata version is written
--> the devices come back online here
4) the mirror device suspend/resume is called to update DM tables
5) during the suspend/resume cycle, *pre*-commit metadata is read;
however, since the failed devices are now back online, we get back
inconsistent set of precommit metadata and the whole operation fails
The patch relaxes the check that fails in step 5 above, namely by ignoring
inconsistencies coming from PVs that are marked MISSING.
This was missing in liblvm and it caused all udev-related operations to
not take effect when using liblvm, e.g. obtaining the list of devices from udev
db instead of scanning the whole /dev which also recreated the .cache as a side
effect. This was also the case with udisks-lvm-pv-export prober which is run
from within udev rules whenever the CHANGE event is fired.
and use this for the LVM critical section logic. Also report an error if
code tries to load a table while any device is known to be in the
suspended state.
(If the variety of problems these changes are showing up can't be fixed
before the next release, the error messages can be reduced to debug
level.)
are affected by the move. (Currently it's possible for I/O to become
trapped between suspended devices amongst other problems.
The current fix was selected so as to minimise the testing surface. I
hope eventually to replace it with a cleaner one that extends the
deptree code.
Some lvconvert scenarios still suffer from related problems.
Patch adds check for stripe not only in direct
LV segment but also in mirror image segment.
This prevents bugs like:
# lvcreate -i2 -l10 -n lv vg_test
# lvconvert -m1 -i1 vg_test/lv
# lvreduce -f -l1 vg_test/lv
WARNING: Reducing active logical volume to 4.00 MiB
THIS MAY DESTROY YOUR DATA (filesystem etc.)
Reducing logical volume lv to 4.00 MiB
Segment extent reduction 9 not divisible by #stripes 2
Logical volume lv successfully resized
# lvremove -f vg_test
Segment extent reduction 1 not divisible by #stripes 2
LV segment lv:0-4294967295 is incorrectly listed as being used by LV lv_mimage_0
Internal error: LV segments corrupted in lv_mimage_0.
Currently some operation with striped mirrors lead
to corrupted metadata, this patch just add detection of such
situation.
Example:
# lvcreate -i2 -l10 -n lvs vg_test
# lvconvert -m1 vg_test/lvs
# lvreduce -f -l1 vg_test/lvs
Reducing logical volume lvs to 4.00 MiB
Segment extent reduction 9not divisible by #stripes 2
Logical volume lvs successfully resized
# lvremove vg_test/lvs
Segment extent reduction 1not divisible by #stripes 2
LV segment lvs:0-4294967295 is incorrectly listed as being used by LV lvs_mimage_0
Internal error: LV segments corrupted in lvs_mimage_0.
We should never remove more extents than requested by user,
so round up to next stripe boundary during lvreduce.
Also this fixes round to zero sized LV bug:
# lvcreate -i2 -I 64k -l10 -n lvs vg_test
# lvreduce -f -l1 vg_test/lvs
Rounding size (1 extents) down to stripe boundary size for segment (0 extents)
WARNING: Reducing active logical volume to 0
THIS MAY DESTROY YOUR DATA (filesystem etc.)
Reducing logical volume lvs to 0
Failed to suspend lvs
There's a possibility someone will use the '/' in the hostname. Since we
generate a temporary file name (path) including the hostname, any '/' would
be ambiguous.
We can always set such hostname using 'sethostname' from unistd.h. But the
'hostname' command already includes the check and removes the '/' char.
However, some old versions still allow that.
See: https://bugzilla.redhat.com/show_bug.cgi?id=711445.
Since this is only a temporary name and the possibility of this error is
quite negligible, we don't need any complex escape sequence here, just a
simple char replace.
Before, we used vg_write_lock_held call to determnine the way a device is
opened. Unfortunately, this opened many devices in RW mode when it was not
really necessary. With the OPTIONS+="watch" rule used in the udev rules,
this could fire numerous events while closing such devices (and it caused
useless scans from within udev rules in return).
A common bug we hit with this was with the lvremove command which was unable
to remove the LV since it was being opened from within the udev rules. This
patch should minimize such situations (at least with respect to LVM handling
of devices).
Though there's still a possibility someone will open a device 'outside' in
parallel and fire the event based on the watch rule when closing a device
once opened for RW.
allocates these buffers in such way it adds memory page for each such buffer
and size of unlock memory check will mismatch by 1 or 2 pages.
This happens when we print or read lines without '\n' so these buffers are
used. To avoid this extra allocation, use setvbuf to set these bufffers ahead.
Signed-off-by: Zdenek Kabelac <zkabelac@redhat.com>
Reviewed-by: Milan Broz <mbroz@redhat.com>
Reviewed-by: Petr Rockai <prockai@redhat.com>
Also, add a new 'obtain_device_list_from_udev' setting to lvm.conf with which
we can turn this feature on or off if needed.
If set, the cache of block device nodes with all associated symlinks
will be constructed out of the existing udev database content.
This avoids using and opening any inapplicable non-block devices or
subdirectories found in the device directory. This setting is applied
to udev-managed device directory only, other directories will be scanned
fully. LVM2 needs to be compiled with udev support for this setting to
take effect. N.B. Any device node or symlink not managed by udev in
udev directory will be ignored with this setting on.
This fails on lenny buildslave for some reason.
For now disable the vgimportclone part of the test until proper fix.
Let the first part of the test still run though, which shows pvs working
with duplicate pvs.
Related to rhbz 697959.
This test fails prior to these two commits:
commit af112eb2c9
Author: Zdenek Kabelac <zkabelac@redhat.com>
Date: Thu Apr 21 13:15:26 2011 +0000
Skip check for NULL before dm_free
dm_free makes this test itself.
commit 91419c3e86
Author: Zdenek Kabelac <zkabelac@redhat.com>
Date: Thu Apr 21 13:13:40 2011 +0000
Fix use of released vgname and vgid
Avoid using of already released memory when duplicated MDA is found.
As get_pv_from_vg_by_id() may call lvmcache_label_scan() use the local copy
of the vgname and vgid on the stack as vginfo may dissapear and code was
then accessing garbage in memory.
i.e. pvs /dev/loop0
(when /dev/loop0 and /dev/loop1 has same MDA content)
Invalid read of size 1
at 0x523C986: dm_hash_lookup (hash.c:325)
by 0x440C8C: vginfo_from_vgname (lvmcache.c:399)
by 0x4605C0: _create_vg_text_instance (format-text.c:1882)
by 0x46140D: _text_create_text_instance (format-text.c:2243)
by 0x47EB49: _vg_read (metadata.c:2887)
by 0x47FBD8: vg_read_internal (metadata.c:3231)
by 0x477594: get_pv_from_vg_by_id (metadata.c:344)
by 0x45F07A: _get_pv_if_in_vg (format-text.c:1400)
by 0x45F0B9: _populate_pv_fields (format-text.c:1414)
by 0x45F40F: _text_pv_read (format-text.c:1493)
by 0x480431: _pv_read (metadata.c:3500)
by 0x4802B2: pv_read (metadata.c:3462)
Address 0x652ab80 is 0 bytes inside a block of size 4 free'd
at 0x4C2756E: free (vg_replace_malloc.c:366)
by 0x442277: _free_vginfo (lvmcache.c:963)
by 0x44235E: _drop_vginfo (lvmcache.c:992)
by 0x442B23: _lvmcache_update_vgname (lvmcache.c:1165)
by 0x443449: lvmcache_update_vgname_and_id (lvmcache.c:1358)
by 0x443C07: lvmcache_add (lvmcache.c:1492)
by 0x46588C: _text_read (text_label.c:271)
by 0x466A65: label_read (label.c:289)
by 0x4413FC: lvmcache_label_scan (lvmcache.c:635)
by 0x4605AD: _create_vg_text_instance (format-text.c:1881)
by 0x46140D: _text_create_text_instance (format-text.c:2243)
by 0x47EB49: _vg_read (metadata.c:2887)
Add testing script
pv_manip.c to properly account for case when pe_start=0 and the first
physical extent is to be released (currently skip the first extent to
avoid discarding the PV label).
My previous patch fixed incorrect error check for dm_snprintf.
However in this particular case - dm_snprintf has been used differently -
just like strncpy + setting last char with '\0' - so the code had to return
error - because the buffer was to short for whole string.
Patch replaces it with real strncpy.
Also test for alloca() failure is removed - as the program behaviour
is rather undefined in this case - it never returns NULL.
lvseg properties for lvm2app, 'devices' and 'seg_pe_ranges'.
Signed-off-by: Dave Wysochanski <dwysocha@redhat.com>
Reviewed-by: Petr Rockai <prockai@redhat.com>
dm_pool_free incorrectly. This check-in fixes that incorrect usage.
I've also added a WHATS_NEW line to reflect the changes I made to allow
lv_extend to operate on 0 length intrinsically layered LVs (i.e mirrors
and RAID). I forgot that in the last commit.
allows us to allocate all images of a mirror (or RAID array) at one
time during create.
The current mirror implementation still requires a separate allocation
for the log, however.
Actually, we can call vg_set_fid(vg, NULL) instead of calling
destroy_instance for all PV structs and a VG struct - it's the same
code we already have in the vg_set_fid.
Also, allow exactly the same fid to be set again for the same PV/VG
Before, this could end up with the fid destroyed because we destroyed
existing fid first and then we used the new one and we didn't care
whether existing one == new one by chance.
As now the FID management is more complex, the code inside free_vg needs
to access some parts of memory pools which were not needed before.
For this - makes the order of unlock_and_free_vg() unconditional.
Keek using unlock_and_free_vg() API function.
For properly working VG locking mechanism only the alphabeting locking
orderer needs to be preserved.
TODO: there could be few more code parts simplified when we 'officially'
support of referencies between different memory pools.
Instead of searching linear list of all LVs, PVs - use created hash tables
also for quick mapping between LV.
(Note - for small number of PVs or LVs the overhead of the hash is bigger).
TODO: Use hash tables in volume_group structure directly.
Instead of regenerating config tree and parsing same data again,
check whether export_vg_to_buffer does not produce same string as
the one already cached - in this case keep it, otherwise throw cached
content away.
For the code simplicity calling _free_cached_vgmetadata() with
vgmetadata == NULL as the function handles this itself.
Note: sometimes export_vg_to_buffer() generates almost the same data
with just different time stamp, but for the patch simplicity,
data are reparsed in this case.
This patch currently helps for vgrefresh.
Align strdup char* allocation just on 2 bytes.
It looks like wasting space to align strings on 8 bytes.
(Could be even 1byte - but for hashing it might eventually get better
perfomance - but probably hardly measurable).
TODO: check on various architectures it's not making any problems.
Isn't usually perfomance critical - but log_error is used i.e.for debuging,
this code noticable slows down the processing.
Added 512KB limit to avoid memory exhastions in case of some endless loop.
TODO: use _lvm_errmsg buffer only when lvm2api needs it.
Avoid locking sum testing with valgrind compilation.
Make memory unaccessible in the valgrind for dm_pool_abadon_object.
Valgrind hinting should not be needed in _free_chunk for dm_free.
'a small step' towards cleaner shutdown sequence.
Normally clvmd doens't care about unreleased memory on exit -
but for valgrind testing it's better to have them cleaned all.
So - few things are left on exit path - this patch starts to remove
just some of them.
1. lvm_thread_fs is made as a thread which could be joined on exit()
2. memory allocated to local_clien_head list is released.
(this part is somewhat more complex if the proper reaction is
needed - and as it requires some heavier code moving - it will
be resolved later.
Could be reached via few of our lvm2 test cases:
==11501== Invalid read of size 8
==11501== at 0x49B2E0: _area_length (import-extents.c:204)
==11501== by 0x49B40C: _read_linear (import-extents.c:222)
==11501== by 0x49B952: _build_segments (import-extents.c:323)
==11501== by 0x49B9A0: _build_all_segments (import-extents.c:334)
==11501== by 0x49BB4C: import_extents (import-extents.c:364)
==11501== by 0x497655: _format1_vg_read (format1.c:217)
==11501== by 0x47E43E: _vg_read (metadata.c:2901)
cut from t-vgcvgbackup-usage.sh
--
pvcreate -M1 $(cat DEVICES)
vgcreate -M1 -c n $vg $(cat DEVICES)
lvcreate -l1 -n $lv1 $vg $dev1
--
Idea of the fix is rather defensive - to allocate one extra element
to 'map' array which is then used in _area_length() - where the
loop checks, whether next map entry is continuous.
By placing there always one extra zero entry -
we fix the read of unallocated memory, and we make sure the data would
not make a continous block.
FIXME: there could be a problem if some special broken lvm1 data would be imported.
As the format1 is currently not really used - leave it for future fix
and use this small hotfix for now.
With recent update of dm_report_field_string() API call to accept
completely const objects - we no longer need loose constness here
and keep it forwarding.
Invalid primary_vginfo was supposed to move all its lvmcache_infos to
orphan_vginfo - however it has called _drop_vginfo() inside the loop
that released primary_vginfo itself - thus made the loop using released
memory.
Use _vginfo_detach_info() instead and call _drop_vginfo after
th loop is finished.
Valgrind trace it should fix:
Invalid read of size 8
at 0x41E960: _lvmcache_update_vgname (lvmcache.c:1229)
by 0x41EF86: lvmcache_update_vgname_and_id (lvmcache.c:1360)
by 0x441393: _text_read (text_label.c:329)
by 0x442221: label_read (label.c:289)
by 0x41CF92: lvmcache_label_scan (lvmcache.c:635)
by 0x45B303: _vg_read_by_vgid (metadata.c:3342)
by 0x45B4A6: lv_from_lvid (metadata.c:3381)
by 0x41B555: lv_activation_filter (activate.c:1346)
by 0x415868: do_activate_lv (lvm-functions.c:343)
by 0x415E8C: do_lock_lv (lvm-functions.c:532)
by 0x40FD5F: do_command (clvmd-command.c:120)
by 0x413D7B: process_local_command (clvmd.c:1686)
Address 0x63eba10 is 16 bytes inside a block of size 160 free'd
at 0x4C2756E: free (vg_replace_malloc.c:366)
by 0x41DE70: _free_vginfo (lvmcache.c:980)
by 0x41DEDA: _drop_vginfo (lvmcache.c:998)
by 0x41E854: _lvmcache_update_vgname (lvmcache.c:1238)
by 0x41EF86: lvmcache_update_vgname_and_id (lvmcache.c:1360)
by 0x441393: _text_read (text_label.c:329)
by 0x442221: label_read (label.c:289)
by 0x41CF92: lvmcache_label_scan (lvmcache.c:635)
by 0x45B303: _vg_read_by_vgid (metadata.c:3342)
by 0x45B4A6: lv_from_lvid (metadata.c:3381)
by 0x41B555: lv_activation_filter (activate.c:1346)
by 0x415868: do_activate_lv (lvm-functions.c:343)
problematic line:
dm_list_iterate_items_safe(info2, info3, &primary_vginfo->infos)
Fix 2 more functions sending cluster messages to avoid passing uninitilised bytes
and compensate 1 extra byte attached to the message from the clvm_header.args[1]
member variable.
If _move_lv_segments is passed a 'lv_from' that does not yet
have any segments, it will screw things up because the code
that does the segment copy assumes there is at least one
segment. See copy code here:
lv_to->segments = lv_from->segments;
lv_to->segments.n->p = &lv_to->segments;
lv_to->segments.p->n = &lv_to->segments;
If 'segments' is an empty list, the first statement copies over
the values, but the next two reset those values to point to the
other LV's list structure. 'lv_to' now appears to have one
segment, but it is really an ill-set pointer.
When I see 'seg_is_mirrored', I expect the argument to be an lv_segment.
In this case, it is lvcreate_params. Both structures, have a 'segtype'
entry which the macro dereferences. However, it just seems easier to
understand if we do 'segtype_is_mirrored' instead.
other cases, the code could wind up removing wrong number of mirrors. In yet
other cases, we could remove the right number of mirrors, but fail to respect
the removal preferences (i.e. keep an image that was requested to be removed
while removing an image that was requested to be kept). Under some
circumstances, remove_mirror_images could also get stuck in an infinite loop.
This patch should fix all of the above undesirable behaviours.
Signed-off-by: Petr Rockai <prockai@redhat.com>
Reviewed-by: Jonathan Brassow <jbrassow@redhat.com>
LVM doesn't behave correctly if running as non-root user,
there is warning when it detects it.
Despite this, it produces many error messages, saying nothing.
See https://bugzilla.redhat.com/show_bug.cgi?id=620571
This patch fixes two things:
1) Removes eror message from device_is_usable() which has no
information value anyway (real warning is printed inside it).
2) it fixes device-mapper initialization, if we support
core dm module autoload and device node is present, it should
fail early and not try recreate existing and correct node.
(non-root == permission denied here)
N.B. In future code should support user roles, some more
drastic checks in code are probably contraproductive now.
Attach \0 for proper char* display - otherwise somewhat random message could
be displayed in debug more and read of unpredictable read of uninitilized
memory values could happen.
As code uses strncpy(system_id, NAME_LEN) and doesn't set '\0'
Fix it by always allocating NAME_LEN + 1 buffer size and with zalloc
we always get '\0' as the last byte.
This bug may trigger some unexpected behavior of the string operation
code - depends on the pool allocator.
FIXME: refactor this code to alloc_vg.
We have 3 components and traling '\0' so allocate proper room for all of them.
Problem was nicely hidden by allocation from pool and allocation aligment
offset - so to trigger real problem with this one is actually hard.
Return value of readlink limits valid string size.
Characters after returned size present some garbage to printf.
Fix it by placing '\0' on the return size value.
Missing free_vg on error_path in lvmcache_get_vg fn. Call destroy_instance
only if the fid is not part of the vg in backup_read_vg fn (otherwise it's
part of the VG we're returning and we definitely don't want to destroy it!).
This is necessary for proper format instance ref_count support. We iterate
over vg->pvs and vg->removed_pvs list and the ref_count is decremented and
then it is destroyed if not referenced anymore.
Since format instances will use own memory pool, it's necessary to properly
deallocate it. For now, only fid is deallocated. The PV structure itself
still uses cmd mempool mostly, but anytime we'd like to add a mempool
in the struct physical_volume, we can just rename this fn to free_pv and
add the code (like we have free_vg fn for VGs).
This is essential for proper format instance ref_count support. We must
use these functions to set the fid everywhere from now on, even the NULL
value!
We'd like to use the fid mempool for text_context that is stored
in the instance (we used cmd mempool before, so the order of
initialisation was not a matter, but now it is since we need to
create the fid mempool first which happens in create_instance fn).
The text_context initialisation is not needed anywhere outside the
create_instance fn so move it there.
Format instances can be created anytime on demand and it contains
metadata area information mostly (at least for now, but in the future,
we may store more things here to update/edit in a PV/VG). In case we
have lots of metadata areas, memory consumption will rise. Using cmd
context mempool is not quite optimal here because it is destroyed too
late. So let's use a separate mempool for format instances.
Reference counting is used because fids could be shared, e.g. each PV
has either a PV-based fid or VG-based fid. If it's VG-based, each PV has
a shared fid with the VG - a reference to VG's fid.
Makes the code more readable and has a smaller number of memory
accesses thus it's small optimisation as well.
For _get_token() optimize number parsing. Check for '.' char only
if it's not a digit. Move pointer incrementation into one place.
For _eat_space() check only p->te for '\0' in skipping of comment line.
Avoid check for '\0' when we know it is space. Also master while loop
doesn't need checking p->tb for '\0'. We just need to check p->tb
isn't already at the end of buffer. This could give 'extra' loop cycle
if we are already there - but safes memory access in every other case.
Add _lv_postorder_vg() - for calling _lv_postorder() for every LV from VG.
We use this in 2 places - vg_mark_partial_lvs() and vg_validate()
so make it as a one function.
Benefit here is - to use only one cleanup code and avoid
potentially duplicate scans of same LVs.
Copy this file as '.gdbinit' to your home directory or your working
directory. It adds the following commands to gdb:
- first_seg
- lv_status
- lv_status_r
- lv_is_mirrored
- seg_item
- seg_status
- segs_using_this_lv
You can get a list of these user-defined commands by typing:
(gdb) help user-defined
You can get more information on each command by typing:
(gdb) help <command>
Accelerate validation loop by using lvname, lvid, pvid hash tables.
Also merge pvl loop into one cycle now - no need to scan the list twice.
List scan is stopped when dm_hash_insert fails.
The error message with loop_counter1 is no longer provided - however
the message has been misleading anyway.
dm_hash binary functions takes void* key - so there is no need to cast
pointers to char* (also the hash key does not have trailing '\0').
This is slight API change, but presents no change for the API user side
it just allows to write code easier as the casting could be removed.
Create new function alloc_vg() to allocate VG structure.
It takes pool_name (for easier debugging).
and also take vg_name to futher simplify code.
Move remainder of _build_vg_from_pds to _pool_vg_read
and use vg memory pool for import functions.
(it's been using smem -> fid mempool -> cmd mempool)
(FIXME: remove mempool parameter for import functions and use vg).
Move remainder of the _build_vg to _format1_vg_read
Fixing few issues:
struct clvm_header contains 'char args[1]' - so adding '+ 1' here
for message length calculation is 1 byte off.
Message with last byte uninitialized is then passed to write function.
Update also related arglen.
Initialise xid and clintid to 0.
Memory allocation is checked for NULL
When the ->params string is empty - memory access is made on the byte
before allocated buffer (catched by valgrind) - in the case it would
constain 0x20 - it would even overwrite this buffer.
So fix by checking len > 0 before doing such access.
Also slightly optimise this loop from repeated strlen call.
lvseg_segtype_dup used memory pool vg memory pool for strind duplication.
However this one gets released before reporting happens so the command like:
pvs -o segtype
prints data from already released memory pool. Thanks to the fact there
is not much allocation happing after the VG is released, the memory
stays unmodified and correct result is printed.
Fix adds support for mempool passed parameter (like other similar
query commands) and uses dm_report memory pool for string duplication.
Failure on pipe kills lv command and produces memory stacktrace.
Add 'true' for empty grep/sed to avoid generating stacktrace.
This should help with long time (7min) for this test on Lenny testing maching
(mimage was filler with stacktrace and each word was used as LV name).
This lvconvert is now able to complete operation automatically without question.
So remove 'echo y |' - it's breaking pipe on some bash versions.
i.e.:
+ echo y
./t-lvconvert-mirror-basic.sh: line 104: echo: write error: Broken pipe
while command has completed this operation succesfully
While STRIPE_SIZE_LIMIT * 2 is basically UINT_MAX, 32bit integer
value can already overflow durin arg size parsing.
(This really happens in test where --stripesize 4294967291 is used,
in s390x uint overflow and this test is ineffective.)
- returned char not needed to be explicitly const
- warn if pipe() fails in clvmd (more fixes here needed for error paths...)
- assign (and ignore) read() output in drain buffer
Currently the test are not properly working with new code.
Use settings for the previous behaviour.
FIXME: update tests to properly test new allocation policies.
We allow writing non-orphan PVs only for resize now. The "orphan PV" assert
in pv_write fn uses the "allow_non_orphan" parameter to control this assert.
However, we should find a more elaborate solution so we can remove this
restriction altogether (pv_write together with vg_write is not atomic, we
need to find a safe mechanism so there's an easy revert possible in case of
an error).
There is a lot to test.
Two new config settings added that are intended to make the code behave
closely to the way it did before - worth a try if you find problems.
Add a small fix that preserves pe_start for lvm1 PVs when being converted.
(this fix needs to be replaced with something more clever, but let's have this working now)
If the PV is already part of the VG (so the pv->fid == vg->fid), it makes no
sense to attach the mdas information from PV to a VG. Instead, we read new
PV metadata information from cache and attach it to the VG fid.
This function also sets a reference to a new VG format instance for all PVs
that are part of the VG so the PV-VG interconnection is consistent after the
change.
Add supporting functions to work with the format instance and metadata area
structures stored within the format instance. Add support for simple indexing
of metadata areas using PV id and mda order (for on-disk PV only for now, we
can extend the indexing even for other mdas if needed - we only need to define
a proper key for the index).
As the kernel seems to be doing weird things during
mlock -> munlock - allow 1 page locking difference without
warning - and log just debug message for a 1 page difference.
Allocation happens outside critical section probably during
log_warn printing.
Should make tests passing for now.
Fixing some const warnings - with API change in:
int vg_extend(struct volume_group *vg, int pv_count, const char *const *pv_names,
Change is needed - as lvm2api expects const behaviour here.
So vg_extend() is doing local strdup for unescaping.
skip_dev_dir return const char* from const char* vg_name.
Rest of the patch is cleanup of related warnings.
Also using dm_report_filed_string() API change to simplify
casting in _string_disp and _lvname_disp.
As dm_report_field_string() doesn't modify content of data pointer,
it can be marked as const.
It's slight API change - but doesn't require any change on the user side
and supports wider range of arguments without const casting.
(i.e. we may use as paramater const lv struct this way: &lv->name)
New strategy for memory locking to decrease the number of call to
to un/lock memory when processing critical lvm functions.
Introducing functions for critical section.
Inside the critical section - memory is always locked.
When leaving the critical section, the memory stays locked
until memlock_unlock() is called - this happens with
sync_local_dev_names() and sync_dev_names() function call.
memlock_reset() is needed to reset locking numbers after fork
(polldaemon).
The patch itself is mostly rename:
memlock_inc -> critical_section_inc
memlock_dec -> critical_section_dec
memlock -> critical_section
Daemons (clmvd, dmevent) are using memlock_daemon_inc&dec
(mlockall()) thus they will never release or relock memory they've
already locked memory.
Macros sync_local_dev_names() and sync_dev_names() are functions.
It's better for debugging - and also we do not need to add memlock.h
to locking.h header (for memlock_unlock() prototyp).
Add configurable option to define minimal size of
of block device usable as a PV.
pv_min_size() is added to lvm-globals and it's being
initialized through _process_config.
Macro PV_MIN_SIZE is unused and removed.
New define DEFAULT_PV_MIN_SIZE_KB is added to lvm-global
and unlike PV_MIN_SIZE it uses KB units.
Should help users with various slow devices attached to the system,
which cannot be easily filtered out (like FDD on /dev/sdX):
https://bugzilla.redhat.com/show_bug.cgi?id=644578
results in clvmd deadlock
When a logical volume is activated exclusively in a cluster, the
local (non-cluster-aware) target is used. However, when creating
a snapshot on the exclusive LV, the resulting suspend/resume fails
to load the appropriate device-mapper table - instead loading the
cluster-aware target.
This patch adds an 'exclusive' parameter to the pertinent resume
functions to allow for the right target type to be loaded.
As functions compiled within this define are apparently stil part of the public API,
(though lvm2 code is never using them unless this define is used for compilation),
keep functions available in the code for now -> revert.
Fix regresion from 2.02.75 speedup - so currently crc32 is a little bit
more complicated on big-endian CPU as the uint32_t needs to be shifted
on here.
Make configurable default behaviour how to deal with device node creates.
With udev system natural options should be 'resume'.
For older systems where user expect there is node in /dev/mapper immediately
after dmsetup create --notable - use 'create'
FIXME:
Code needs fixing passing this flag through udev cookie.
activated.
In order to achieve this, we need to be able to query whether
the origin is active exclusively (a condition of being able to
add an exclusive snapshot).
Once we are able to query the exclusive activation of an LV, we
can safely create/activate the snapshot.
A change to 'hold_lock' was also made so that a request to aquire
a WRITE lock did not replace an EX lock, which is already a form
of write lock.
Add new function dm_task_set_add_node() to select between 2 types
of node creation in device directory.
DM_ADD_NODE_ON_RESUME is now default and ensures node is created on
resume. Old original behavior is accessible with DM_ADD_NODE_ON_CREATE.
In this case node would be created during dmsetup create --notable.
For the user 2 new options for dmsetup create are added:
[{--addnodeonresume | --addnodeoncreate }]
Properly working node creation on resume is needed for proper operation
stacking and ability to correctly check in which state the device should
after whole udev transation.
Remove temporaly added fs_unlock() calls to fix clmvd usablity.
Now when the message passing is properly working - they are no longer needed.
Simplify no_locking check for VG unlock - as message is always send
for all targets - clustered & non-clustered.
Thanks to CLVMD_CMD_SYNC_NAMES propagation fix the message passing started
to work. So starts to send a message before the VG is unlocked.
Removing also implicit sync in VG unlock from clmvd as now the message
is delievered and processed in do_command().
Also add support for this new message into external locking
and mask this event from further processing.
With the ability to stack many operations in one udev transaction -
in same cases we are adding and removing same device at the same time
(i.e. deactivate followed by activate).
This leads to a problem of checking stacked operations:
i.e. remove /dev/node1 followed by create /dev/node1
If the node creation is handled with udev - there is a problem as
stacked operation gives warning about existing node1 and will try to
remove it - while next operation needs to recreate it.
Current code removes all previous stacked operation if the fs op is
FS_DEL - patch adds similar behavior for FS_ADD - it will try to
remove any 'delete' operation if udev is in use.
For FS_RENAME operation it seems to be more complex. But as we
are always stacking FS_READ_AHEAD after FS_ADD operation -
should be safe to remove all previous operation on the node
when udev is running.
Code does same checking for stacking libdm and liblvm operations.
As a very simple optimization counters were added for each stacked ops
type to avoid unneeded list scans if some operation does not exists in
the list.
Enable skipping of fs_unlock() (udev sync) if only DEL operations are staked.
as we do not use lv_info for already deleted nodes.
This is better way how to fix clustered synchronization with udev.
As the code for message passing needs fixed - put currently
fs_unlock() after every active/deactive command in clvmd to
ensure nodes are properly created in time.
It would be most useful to add "dmsetup ls --tree" to the commands run.
This command helps in answering the question "which devices are actually
underneath a given LV?"
Although the info is available with other existing dmsetup commands,
adding this command gives a much clearer summary of complex setups.
Here's an example of an LVM mirror, with mirror images on partitions
created on top of multipath devices. The multipath devices are on
simple block devices. As you can see, it is easy to see the stacking
from the "dmsetup ls --tree" output:
vgmpathtest-lvmpathmir (253:14)
├─vgmpathtest-lvmpathmir_mimage_1 (253:13)
│ └─mpath5p1 (253:5)
│ └─mpath5 (253:2)
│ ├─ (8:16)
│ └─ (8:0)
├─vgmpathtest-lvmpathmir_mimage_0 (253:12)
│ └─mpath6p1 (253:6)
│ └─mpath6 (253:3)
│ ├─ (8:48)
│ └─ (8:32)
└─vgmpathtest-lvmpathmir_mlog (253:11)
└─mpath7 (253:4)
├─ (8:80)
└─ (8:64)
VolGroup00-LogVol01 (253:1)
└─ (202:2)
vgtest-lvmir (253:10)
├─vgtest-lvmir_mimage_1 (253:9)
│ └─ (7:1)
├─vgtest-lvmir_mimage_0 (253:8)
│ └─ (7:0)
└─vgtest-lvmir_mlog (253:7)
└─ (7:3)
VolGroup00-LogVol00 (253:0)
└─ (202:2)
But it is much harder to see the stacking with only the commands today
("dmsetup info", "dmsetup status", and "dmsetup table"). We could
piece together the stacking from "dmsetup table" but it requires
further processing (take output from "dmsetup info to get
map name to major/minor, then parse "dmsetup table", etc).
There was no effect from having this wrong yet, because the
tree of callers only ever cared about the answer to the first
condition (!response), which determines whether a lock is
held or not. Correct responses, however, are needed soon.
Instead of implicitly syncing udev operation in clustered and
file locking code - call synchronization directly in lock_vol() when
the operation unlocks VG
The problem is missing implicit fs_unlock() in the no_locking code.
This is used with --sysinit on read-only filesystem locking dir.
In this case vgchange -ay could exit before all udev nodes are properly
synchronised and may cause problems with accessing such node right after
vgchange --sysinint command is finished.
Add test case for vgchange --sysinit.
As the option 'set -e -o pipefail' is very sensite on pipe breaking
stop using '-q' for grep commands.
Otherwise this command (with large enough table) would fail:
dmsetup table | egrep -q
with exit code 141 (128 + SIGPIPE)
As Peter suggested, he prefers to keep '-o pipefail' - so make sure all
piped commands will read the whole output and will not exit too early.
This is to avoid any scanning and processing of DM devices while they are in
suspended state (e.g. a rename while the device is suspended - a CHANGE event
is generated!). Otherwise, any scanning in the rules could end up with locking
the calling process until the device is resumed and so we don't receive a
notification about udev rules completion until then (and that effectively
locks out the process awaiting the notification!).
However, we still keep 'disk' and any 'subsystem' related udev rules running.
We trust these and these should check themselves whether a device is suspended
or not, not trying to run any scanning if it is.
strncpy (which check each byte for \0) is not need as we always copy
the length size - so using memcpy is a bit cheaper.
Add missing log_error message for failed allocation.
Fix assert abort of dmsetup (when compiled with pool debug)
dmsetup splitname --nameprefixes --noheadings --rows gvg-a2
Move pool begin in the inner loop - otherwise it would using
already 'ended' pool object.
return lockspace reference (even if lockspace already exists)
and thus increases DLM lockspace count. It means that after
clvmd restart the lockspace is still in use.
(The only way to clean environment to enable clean cluster
shutdown is call "dlm_tool leave clvmd" several times.)
Because only one clvmd can run in time, we can use simpler logic,
try to open lockspace with dlm_open_lockspace() and only if it fails
try to create new one. This way the lockspace reference doesn not
increase.
Very easily reproducible with "clvmd -S" command.
Patch also fixes return code when clvmd_restart fails and fixes
double free if debug option was specified during restart.
Fixes https://bugzilla.redhat.com/show_bug.cgi?id=612862
Improve condition within lock_vol so we are not calling extra unlock
if the volume just has been deactivated.
Patch uses lck_type and replaces negative 'and' condition to more
readable 'or' condition.
Few missing strace traces added.
As sync_local_dev_names() cannot be called within activation context,
add new parametr which allows to select if the sync call is needed
before executing new command.
Instead of checking $LOCAL_CLVMD set during some 'aux' execution which
doesn't seem to be propagated to this shell - check for existance of pid
file of clvmd process - so this test is skipped in singlenode cluster test.
Harness seems to be able to busyloop in while cycle and not moving forward
for certain buffer - so check whethere there was some progress.
This fix allows to continue after failed cluster test.
Fix gcc warning for hiding global variable 's' -> sig.
Stop calling fs_unlock() from lv_de/activate().
Start using internal lvm fs cookie for dm_tree.
Stop directly calling dm_udev_wait() and
dm_tree_set/get_cookie() from activate code -
it's now called through fs_unlock() function.
Add lvm_do_fs_unlock()
Call fs_unlock() when unlocking vg where implicit unlock solves the
problem also for cluster - thus no extra command for clustering
environment is required - only lvm_do_fs_unlock() function is added
to call lvm's fs_unlock() while holding lvm_lock mutex in clvmd.
Add fs_unlock() also to set_lv() so the command waits until devices
are ready for regular open (i.e. wiping its begining).
Move fs_unlock() prototype to activation.h to keep fs.h private
in lib/activate dir and not expose other functions from this header.
Change function import_vg_from_buffer() to import_vg_from_config_tree().
Instead of creating config tree inside the function allow config tree to
be passed as parameter - usable later for caching.
If some allocation for peristent filter fails its memory reference
was lost, fix it by calling filter's destructor.
Fix log_error messages for failing allocation.
Variable 'ret' assigned from _do_event() was actually not used and replaced with next
assignment without any read of the returned value.
Code is reformated - so the error path is put in the if() branch and normal
code is put after the 'if' together with FIXME comment.
FIXME lowprio: logging needs to be fixed in this code,
- multiple log_errors are printed, stacks are missing...
a mirror image (or removing a log while adding a mirror). Advise the
user to use two separate commands instead.
This issue become especially problematic when PVs are specified, as they
tend to mean different things when adding vs removing. In a command that
mixes adding and removing, it is impossible to decern exactly what the
user wants.
This change prevents bug 603912.
- somewhat neater, more consistent and more readable output
- possible to set any lvm.conf value: aux lvmconf "section/key = value"
- LVM_TEST_NODEBUG to suppress the (lengthy) "## DEBUG" output
- back-substitution on test output ($TESTDIR/$PREFIX -> @TESTDIR@/@PREFIX@)
- support code moved from test/ to test/lib/ --> less clutter
Checking for vg being != NULL in this place is not needed.
Pointer vg is already dereferced in this function above this code line.
Also this internal function _read_pv is always called with valid 'vg' pointer.
Call for pthread_join() does not set errno value even though return values
looks like that. For now assign errno from return value and still use
strerror() to print some error message as this seems to be commonly used.
Add also log_sys_error() message for error close of local pipe.
Add checks for clonning allocation a fail-out when something is
not allocated correctly.
Also move var declaration to the begining of the function
and fix log_error messages.
As 'const' types are also passed to macro dm_list_struct_base -
keep offset calculation with const char pointers.
Fixes several gcc constness warnings.
As const segment_type or const format_type are never released
use their non-const version and remove const downcast from dm_free calls.
This change fixes many gcc warnings we were getting from them.
Change API interface to accept even completely const array patterns.
This should present no change for libdm users and allows to pass
pattern arrays without cast to const char **.
To have better control were the config tree could be modified use more
const pointers and very carefully downcast them back to non-const
(for config tree merge).
As ternary operator has lower priority then add operation, this check
was not doing what seemed to be expected.
So enclose the test in braces and check for NULL in *buf.
We need to be sure that /var/run and /var/lock is always there.
(E.g. these two directories could be using tmpfs which then loose
all the content after reboot.)
Detect existence of new SELinux selabel interface during configure.
Use new dm_prepare_selinux_context instead of dm_set_selinux_context.
We should set the SELinux context before the actual file system object creation.
The new dm_prepare_selinux_context function sets this using the selabel_lookup
fn in conjuction with the setfscreatecon fn. If selinux/label.h interface
(that should be a part of the selinux library) is not found during configure,
we fallback to the original matchpathcon function instead.
Set cmd->independent_metadata_areas if metadata/dirs or disk_areas in use.
- Identify and record this state.
Don't skip full scan when independent mdas are present even if memlock is set.
- Clusters and OOM aren't supported, so no problem doing the proper scans.
Avoid revalidating the label cache immediately after scanning.
- A simple optimisation.
Support scanning for a single VG in independent mdas.
- Not used by the fix but I left it in anyway as later patches might use it.
This reset of vgmem pointer causes access of already released memory.
(_vg_make_handle allocates vg from vgmem pool itself - which is a bit tricky)
Interestingly this memory fault was missed by our test suite.
Add log_error message for lv_info failure and exit from futher
processing.
Replace 'leg' occurence in debug message with 'image' which
is used in other messages.
Patch adds extra check for lv_name not being NULL.
Test avoids unneeded strlen call for this case.
Otherwise there is no functional change as test would fail on
size_t comparation even for NULL lv_name (thus there is no risk
of NULL dereference when taking 'true' if branch.
Add test for NULL before passing uuid as src argument to memcpy.
As memcpy function is declared as function not accepting NULL.
Though we pass NULL only with zero length so this patch presents
no functional change to the code.
Add test for NULL from dm_poll_create.
Reorder dm_pool_destroy() before file close and add label out:.
Avoid leaking file descriptor if the allocation fails.
Nicely hidden memory leak in outf macro error path.
This macro is using out_text() and does automagical return_0.
That would leak tag_buffer allocated memory.
As there was same code for tags output - create _out_tags() function.
Set vg to NULL after releasing it as the following memlock() test may
lead to goto for the second call of vg_release() with the already
released vg pointer.
LCK_CACHE is defined as 0x100 so it cannot be passed through
unsigned char parameter - remove it from the sprintf code.
If the LCK_CLUSTER should be printed here - lot of code need
to be reworked - so adding FIXME comment.
was lacking the (vgmem) pool. We now create that pool. There is at least one
more such VG (_dummy_vg) which is pool-less. I am not sure what is the right
way to go about this, but this is currently necessary to fix a segfault
introduced by using vgmem in the reporter in Dave's lvseg lvm2app patches.
Signed-off-by: Petr Rockai <prockai@redhat.com>
Similar to 'get' property internal functions.
Add specific 'set' function for vg_mda_copies.
Signed-off-by: Dave Wysochanski <wysochanski@pobox.com>
Reviewed-by: Petr Rockai <prockai@redhat.com>
Fix for the last commit as $MOUNTED is not only used as bool flag,
but also store mounted location for remount - so parsing output
from mount differently then from /proc/mounts.
Prefix calls of 'tunefs' tools with LANG=C to be sure we always do get
some nonlocalized strings.
Avoid using forced 'resize2fs' for cleanly unmounted filesystems and
run regular fsck -f for this case as required by resize2fs.
'fsadm check' uses date difference for extX filesystems between
the last mount and last check of 'fsck -f' execution and if the mount
was later run 'fsck' with -f so resize2fs is happy and user does not
need to pass '-f' flag.
As util-linux package seems to give all the time different names,
try harder to figure out, where is the given lv possible mounted
and scan /proc/mounts and if not found there, test also 'mount' output.
/dev/dm-xxx
/dev/mapper/vg-lv
/dev/vg/lv
All of them could be used different combination in /proc/mount and mount output.
Patch fixes regression for older systems where new detection code failed to
find valid combination.
Fixing warning introduced by 'include make.tmpl' commit.
Produced this warning:
Makefile:29: warning: overriding commands for target `distclean'
../make.tmpl:366: warning: ignoring old commands for target `distclean'
Updated patch from Florian Haas from Linux-HA project.
User needs to 'configure --enable-ocf' to get file installed
by 'make install' target by default.
User can also use 'make install_ocf' to get only ocf files installed.
With disabled (default) ocf support - no ocf files are installed.
FIXME: ocf installation path needs to be kept in sync with pacemaker.
find better way and possible also better location.
Patch updates exec_cmd() and adds 3rd parameter with pointer for
status value, so caller might examine returned status code.
If the passed pointer is NULL, behavior is unmodified.
Patch allows to confinue with lvresize if the failure from fsadm check is
caused by mounted filesystem as many of filesystem resize tools do support
online filesystem resize. (originally user had to use flag '-n' to bypass
this filesystem check)
Return status code 3 for fsadm check of mounted filesystem - used later with
lvresize update patch to better support online filesystem resize.
Also makes a more consistent user interruption and returns status code 2
in this case.
Fix memory leak of field_id in _output_field function.
There's been a patch added recently to use dynamic allocation for metadata
tags buffer to remove the 4k limit (for writing metadata out). However, when
using reporting commands like vgs and lvs, we still need to fix libdm reporting
functions themselves to support such long outputs. So the buffer used in those
reporting functions is dynamic now.
The patch also includes a fix for field_id memory leak which was found in
the _output_field function.
The management threads (main_loop, the socket thread) could close a single fd
twice in a row sometimes. At least one other thread can be running at the same
time as the threads doing the double close. That one running thread also
happens to do some IO (namely, open /proc/devices, read from it, close it). If
there was enough "demand" for the local socket, this could happen:
- a connection to clvmd is about to finish, let's say the fd is 13 (it often
happens to be in my test script, don't ask why)
- the local_sock thread calls close(13)
- the lvm thread calls open("/proc/devices"...) and gets 13
- the main_loop thread calls close(13) [OOPS!]
- new connection arrives, and is accept'd by a (new) local_sock thread
- the accept gives an fd of 13 (since it's the lowest free fd at this point)
- the lvm thread gets around to read from it's /proc/devices handle... 13,
again
- the lvm thread hangs forever trying to read from the socket instead of
/proc/devices
Signed-off-by: Petr Rockai <prockai@redhat.com>
Reviewed-by: Milan Broz <mbroz@redhat.com>
Function pull_stateo() checks for NULL 'buf' - but return for this error
path was missing. cmirror code never calls this function with NULL 'buf',
so this fix has no effect on current code base, but makes clang happier.
It's quite new feature which is not supported by older compilers.
So until some better macros are introduced into LVM code - hotfix current
compilation problems and compile this code only for __clang__ defining compilers.
Simultaneous -a and --refresh is not valid.
poll+monitor are valid together with or without -ay* (but not with -an*)
No longer print polling results summary if no LVs in the VG were polled.
Add a generic LV property function to lvm2app, similar to VG function.
Return lvm_property_value and require caller to check 'is_valid' flag
and lvm_errno() for API error.
Add a generic PV property function to lvm2app, similar to VG function.
Return lvm_property_value and require caller to check 'is_valid' flag
before using the value. If 'is_valid' is not set, then lvm_errno()
should be used to obtain the specific error.
Add a generic VG property function to lvm2app. Call the internal library
vg_get_property() function. Strings are dup'd internally.
Rework lvm_vg_get_property to return lvm_property_value and require caller
to check 'is_valid' flag. If !is_valid, the caller can check lvm_errno()
for the specific error.
Create a 'get_property' function, local to lvm2app, that factors out
most of the common code that copies the components of lvm_property_type
into lvm_property_value. This allows for a 1-line function for each
of the generic property functions exported by lvm2app.
Makes clang happier as it covers all code paths and avoids NULL pointer
dereference through the 'com' pointer (which is NULL by default static
initialisation).
Reported by clang as: Argument with 'nonnull' attribute passed null
Reuse the result of the last strchr() call - make sure, 'st' point is not
null for the next strchr() call.
We cast (char*) to (uint32_t*) that changes alignment requierements.
For our case the code has been correct as alloca() returns properly
aligned buffer, however this patch make it cleaner and more readable
and avoids warning generation.
A merged snapshot's DM device is made to use the "error" target as part
of lvm's transaction to merge a snapshot. This snapshot merge use-case
aside, any device using the error target shouldn't be scanned.
Based on review comments, rename a few fields in lvm_property_type.
In particular, change 'is_writeable' to 'is_settable', which is
more intuitive to the intent of the bitfield (a 'set' function
exists for this field/property). Also, remove the char array
for 'id' - unnecessary as we can just use the string passed in
to do the strcmp. Finally rename the union members from n_val
to 'integer' and 's_val' to 'string'.
The signalling code (pthread_cond_signal/pthread_cond_wait) in the
pre_and_post_thread was using the wait mutex (see man pthread_cond_wait)
incorrectly, and this could cause clvmd to deadlock when timing was
right. Detailed explanation of the problem follows.
There is a single mutex around (L for Lock, U for Unlock), a signal (S) and a
wait (W). C for pthread_create. Time flows from left to right, each arrow is a
thread.
So first the "naive" scenario, with no mutex (PPT = pre_and_post_thread, MCT =
main clvmd thread; well actually the thread that does read_from_local_sock). I
will also use X, for a moment when MCT actually waits for something to happen
that PPT was supposed to do.
MCT -----C ------S--X-----S----X----------------------S------XXXXXXXXX
| everything OK up to this --> <-- point...
PPT -----WWW-----WWWW------------------------------WWWWWWWWWWWWW
Ok, so pthread API actually does not let you use W/S like that. It goes out of
its way to tell you that you need a mutex to protect the W so that the above
cannot happen. *But* if you are creative and just lock around the W's and S's,
this happens:
MCT ----C-----LSU----X-----LSU----X------------LSU------XXXXXXX
|
PPT ---LWWWU-------LWWWWU-----------------------LWWWWWWWWW
Ooops. Nothing changed (the above is what actually was done by clvmd before
this satch). So let's do it differently, holding L locked *all* the time in
PPT, unless we are actually in W (this is something that the pthread API does
itself, see the man page).
MCT ----C-----LSU------X---LSU---X-----LLLLLLLSU----X----
| (and they live happily ever after)
PPT L---WWWWW---------WWWW----------------W----------
So W actually ensures that L is unlocked *atomically* together with entering
the wait. That means that unless PPT is actually waiting, it cannot be
signalled by MCT. So if MCT happens to signal it too soon (it wasn't waiting
yet), it (MCT) will be blocked on the mutex (L), until PPT is actually ready to
do something.
to lvm.conf in the activation section: 'snapshot_autoextend_threshold' and
'snapshot_autoextend_percent', that define how to handle automatic snapshot
extension. The former defines when the snapshot should be extended: when its
space usage exceeds this many percent. The latter defines how much extra space
should be allocated for the snapshot, in percent of its current size.
Reorder linked libraries so we better support --as-needed linker flag used
by some distributions (i.e. Gentoo).
Patch suggested by Diego Elio Pettenò <flameeyes <at> gmail.com>
Problem:
When both legs of a mirrored log fail, neither the log nor the parent
mirror can proceed. The repair code must be careful to replace the
log with an error target before operating on the parent - otherwise,
the parent can get stuck trying to suspend because it can't push through
any writes. The steps to replace the log device with an error target
were incomplete and resulted in the replacement not happening at all!
The code originally had all the necessary logic to complete the
replacement task, but was pulled out in a effort to clean-up that
section of code, while fixing another bug:
<offending commit msg>
In addition, I added following three changes.
- Removed tmp_orphan_lvs handling procedure
It seems that _delete_lv() can handle detached_log_lv properly
without adding mirror legs in mirrored log to tmp_orphan_lvs.
Therefore, I removed the procedure.
- Removed vg_write()/vg_commit()
Metadata is saved by vg_write()/vg_commit() just after detached_log_lv
is handled. Therefore, I removed vg_write()/vg_commit().
</offending commit msg>
http://sources.redhat.com/cgi-bin/cvsweb.cgi/LVM2/lib/metadata/mirror.c?cvsroot=lvm2&f=h#rev1.130
I've reverted the "clean-up" changes associated with that fix, but not what
that commit was actually fixing.
Signed-off-by: Jonathan Brassow <jbrassow@redhat.com>
Reviewed-by: Petr Rockai <prockai@redhat.com>
page.
Add ->target_name to segtype_handler to allow a more specific target
name to be returned based on the state of the segment.
Result of trying to merge a snapshot using a kernel that doesn't have
the snapshot-merge target:
Before:
# lvconvert --merge vg/snap
Can't expand LV lv: snapshot target support missing from kernel?
Failed to suspend origin lv
After:
# lvconvert --merge vg/snap
Can't process LV lv: snapshot-merge target support missing from kernel?
Failed to suspend origin lv
Unable to merge LV "snap" into it's origin.
Use _even_rand() function instead of floor() in _bitset_with_random_bits().
floor() function is missing in dietlibc (on architectures other than x86).
Moreover using floor() to clip rand results does not assure even result
distribution. _even_rand() uses integer arithmetic only and is designed to
return evenly distributed results.
> Looks OK to me. It took a while to decipher what is the exact meaning of
> the loop in _even_rand (to a non-pseudorandomness-expert) but I am
> fairly comfortable with it now. If I understand this correctly, it
> rejects numbers that come from an "incomplete" slice of the RAND_MAX
> space (considering the number space [0, RAND_MAX] is divided into some
> "max"-sized slices and at most a single smaller slice, between [n*max,
> RAND_MAX] for suitable n -- numbers from this last slice are discarded
> because they could distort the distribution in favour of smaller
> numbers).
Signed-off-by: Przemyslaw Iskra <sparky <at> pld-linux.org>
Reviewed-by: Petr Rockai <prockai <at> redhat.com>
re-add a physical volume that has gone missing previously, due to a transient
device failure, without re-initialising it.
Signed-off-by: Petr Rockai <prockai@redhat.com>
Reviewed-by: Alasdair Kergon <agk@redhat.com>
Try to distinguish between the case of using interactive shell and non
interactive running - different combinations of '-y' and '-p' option
needs to be used for fsck.
Update the way how fsadm detects mounted filesystem.
With udev /dev/dm-XXX paths are now returned - but mount or /proc/mounts
prints names in form of /dev/mapper/vg-lv - so the match was not found.
Fixex RHBZ #638050.
Current solution uses same trick as mount and detects vg-lv name through
/sys where available - this should be reasonable safe.
Instead of calling mount without parameter to get actual mount table,
switch to use /proc/mounts directly.
Fix missing 'dry' execution of lvresize - fixing problem where resize
command were 'dry-run' executed - but lvresize has been executed for real.
Also adapt code slightly to support better recursive execution of fsadm
through lvresize call.
Under certain conditions it was possible to break (^C) fsadm before actually
resizing filesystem, but lvresize which executed fsadm will think resize
was succesful and shrinks partitions with unresized filesystem on it.
Fix by returning error (1) for this case - this stops lvresize from futher
proceding in resize operation.
In other LVM memory structures such as volume_group, the field
used to store flags is called "status", and on-disk fields are called
'flags', so rename the one inside metadata_area to be consistent.
Not only is it more consistent with existing code but is cleaner
to say "the status of this mda is ignored".
Background for this patch - prajnoha pinged me on IRC this morning
about a fix he was working on related to metadataignore when
metadata/dirs was set. I was reviewing my patches from this year
and realized the 'flags' field was probably not the best choice
when I originally did the metadataignore patches.
Current lvm1 allocation code seems to not properly
map segments on missing PVs.
For now disable this functionality.
(It never worked and previous commit just introduced segfault here.)
So the partial mode in lvm1 can only process missing PVs
with no LV segments only.
Also do not use random PV UUID for missing part but use fixed
string derived from VG UUID (to not confuse clvmd tests).
We need to use a similar function for pv and lv properties, so just make
a generic _get_property() function that contains most of the required
functionality. Also, add a check to ensure the field name matches the
object passed in by re-using report_type_t enum. For pv properties,
the report_type might be either PVS or LABEL.
In addition, add 'const' to 'get' functions object parameter, but not
'set' functions. Add _not_implemented_set() and _not_implemented_get()
functions.
Add 'get' functions based on generic macros for VG, PV, and LV.
Add 'get' functions for vg string fields, vg_name, vg_fmt, vg_sysid,
vg_uuid, vg_attr, and vg_tags, and all numeric fields.
Add supporting functions for vg_name, vg_fmt, vg_system_id.
Append "_dup" to end of supporting functions to make clear the strings
are dup'd and to avoid namespace conflict with vg_name.
Add supporting functions for pv_uuid, vg_uuid, and lv_uuid.
Call new function id_format_and_copy. Use 'const' where appropriate.
Add "_dup" suffix to indicate memory is being allocated.
Call {pv|vg|lv}_uuid_dup from lvm2app uuid functions.
This patch addresses code review request to simplify creation of 'attr'
strings. The simplification is done in this separate patch to more
easily review and ensure the simplification is done without error.
Move the creating of the 'attr' strings into a common function so
they can be called from the 'disp' functions as well as the new
'get' property functions.
Add "_dup" suffix to indicate memory is allocated.
Refactor pvstatus_disp to take pv argument and call pv_attr_dup().
This patch is similar to the other patches for pv and vg
functionality, and separates lv functionality into separate
files, concentrating on reporting fields and simple functions.
The metadata.[ch] files are very large. This patch makes a first
attempt at separating out pv functions and data, particularly
related to the reporting fields calculations.
More code could be moved here but for now I'm stopping at reporting
functions 'get' / 'set' functions.
The metadata.[ch] files are very large. This patch makes a first
attempt at separating out vg functions and data, particularly
related to the reporting fields calculations.
Read complete content of /proc/self/maps into one buffer without
realocation in the middle of reading and before doing any m/unlock
operation with these lines - as some of them gets change.
With previous implementation we've read some mappings twice ([stack])
If some lvm1 device is missing, lvm fails on all operations
# vgcfgbackup -f bck -P vg_test
Partial mode. Incomplete volume groups will be activated read-only.
3 PV(s) found for VG vg_test: expected 4
PV segment VG free_count mismatch: 152599 != 228909
PV segment VG extent_count mismatch: 152600 != 228910
Internal error: PV segments corrupted in vg_test.
Volume group "vg_test" not found
Allow loading of lvm1 partial VG by allocating "new" missing PV,
which covers lost space. Also this fake mising PV inform code
that it is partial VG.
https://bugzilla.redhat.com/show_bug.cgi?id=501390
Revert to old glibc behaviour for vsnprintf used in emit_to_buffer fn.
Otherwise, the check that follows would be wrong for new glibc versions.
This caused the rh bug #633033 to be undetected and pass throught the check,
corrupting the metadata!
In certain configurations, we're not under a VG rw lock while trying to write
a new archive file with VG metadata. A common example is using "vgs" while
having the content of backup and archive directories empty. The code scans the
content of these directories and tries to determine the final index that should
be used in archive name. Since we're not under a lock, we can get into a race
while choosing the index which could end up showing errors about not being able
to rename to final archive name. Let's add random number suffix to these archive
file names so we can avoid the race.
For example, when using '--config "backup { ... }"' line, the values from
lvm.conf (or default values) should be overridden. This patch adds
reinitialisation of archive and backup handling on toolcontext refresh
which makes these settings to be applied.
can be opprobriously slow if created with '--nosync'.
One of the ways cluster mirrors coordinate I/O and recovery
amoung the different machines is by the use of the log
function 'is_remote_recovering()' which lets nodes know if
a region they wish to perform a write on is currently being
recovered on another node. If the region is being recovered,
the I/O is delayed.
The 'is_remote_recovering' routine has been optimized to
avoid the deluge of requests that would be issued to the
userspace log server by maintaining a marker of how far
the recovery has gotten. It can then immediately return
'not recovering' if the region being inquired about is
less than this mark. Additionally, if the region of
concern is greater than the mark, the function will
limit the number of transmissions to userspace by assuming
the region /is/ being recovered when skipping the
transmission. This limits the amount of processing
and updates the mark in 1/4 sec time steps.
This patch fixes a problem where 'the mark' is not being
updated because of faulty logic in the userspace log
daemon. When '--nosync' is used to create a cluster
mirror, the userspace log daemon never has a chance
to update the mark in the normal way. The fix is to set
the mark to "complete" if the mirror was created with
the --nosync flag.
For now force it in lvm.conf (otherwise it fails on older systems like RHEL5
where are these options disabled by default).
FIXME: it should test and detect both versions.
may never complete.
If you convert from a linear to a mirror and then convert that
mirror back to linear /while/ the previous (up)convert is
taking place, the mirror polling process will never complete.
This is because the function that polls the mirror for
completion doesn't check if it is still polling a mirror and
the copy_percent that it gets back from the linear device is
certainly never 100%.
The fix is simply to check if the daemon is still looking at
a mirror device - if not, return PROGRESS_CHECK_FAILED.
The user sees the following output from the first (up)convert
if someone else sneaks in and does a down-convert shortly
after their convert:
[root@bp-01 ~]# lvconvert -m1 vg/lv
vg/lv: Converted: 43.4%
ABORTING: Mirror percentage check failed.
to block when a mirror under a snapshot suffers a failure.
The problem has to do with label scanning. When a mirror suffers
a failure, the kernel blocks I/O to prevent corruption. When
LVM attempts to repair the mirror, it scans the devices on the
system for LVM labels. While mirrors are skipped during this
scanning process, snapshot-origins are not. When the origin is
scanned, it kicks up I/O to the mirror (which is blocked)
underneath - causing the label scan (an thus the repair operation)
to hang.
This patch simply bypasses snapshot-origin devices when doing
labels scans (while ignore_suspended_devices() is set). This
fixes the issue.
introduced in commit b16b4d92a7
"Improve various log messages."
fixes a lot of
../include/metadata.h:148: warning: type qualifiers ignored on function return type
set appropriate Required-Start and Required-Stop at configure time.
Reorder the checks for user selected cluster managers to match auto
detected ones, to be consistent in the output.
Add special case for qdiskd that´s started after cman/lock_gulmd for
RHEL-4/RHEL-5.
If pvmove crashed and metadata contains pvmove LV
but without miorrored segments, pvmove --abort
will not repair the situation (and finish wth success!).
Fix it by allowing metadata update if aborting
(thus removing pvmove LV) even if no moved LVs detected.
(Tested on real metadata provided by an lvm user:-)
detected alignment.
NOTE: lvm2 doesn't detect MD 1.2 metadata (now the default on RHEL6) so
for now I'm forcing 1.0 metadata. This was needed to be able to reuse
the existing loop devices but recreate the md device with different
raid0 striping.
Add "devices/default_data_alignment" to lvm.conf to control the internal
default that LVM2 uses: 0==64k, 1==1MB, 2==2MB, etc.
If --dataalignment (or lvm.conf's "devices/data_alignment") is specified
then it is always used to align the start of the data area. This means
the md_chunk_alignment and data_alignment_detection are disabled if set.
(Same now applies to pvcreate --dataalignmentoffset, the specified value
will be used instead of the result from data_alignment_offset_detection)
set_pe_align() still looks to use the determined default alignment
(based on lvm.conf's default_data_alignment) if the default is a
multiple of the MD or topology detected values.
Add 'get' functions based on the simple macro function definition for a
numeric property.
Add 'get' functions for the following: _vg_extent_count_get,
_vg_free_count_get, _max_lv_get, _max_pv_get, _pv_count_get,
_lv_count_get, _snap_count_get, _vg_seqno_get, _vg_size_get,
_vg_free_get, vg_mda_*.
For size functions, multiply by SECTOR_SIZE to return the value in bytes.
Extend the existing reporting infrastructure definitions and structures
to include a 'get' and 'set' function for each field. We will provide
a 'get' and 'set' function for each of these fields, which will be utilized
by exported lvm2app functions.
Define a default _not_implemented 'get' and 'set' function that just sets
an errno and returns 0. Future patches will actually implement the
specific 'get' and 'set' functions for each property. For read-only
properties, only the 'get' function will be implemented.
Define vg_get_property() function to query a property. We will call
this from a lvm2app function.
Rather than hard code the size of the field, use a #define, so we can re-use.
The #define will be needed in a future patch when we extend the reporting
infrastructure to have 'get' and 'set' functions for each field, allowing
lvm2app functions which query any report field. In order to provide a
generic lookup based on the field id, we will define a type containing this
field id, and thus, we will need to re-use the length of this string as
it's defined inside libdevmapper.h.
The 'id' entries in columns.h are the report field names. Since these are
unique, we'd like to use them in generation of 'get' / 'set' functions.
As a step towards using them for this purpose, remove the explicit double
quotes and use the macro '#' character to add the double quotes back when
placing them into the '_fields' array 'id' member.
Add a 'flags' field to columns.h, and set it to 0 by default.
Define FIELD_MODIFIABLE flag to indicate whether a 'set' function exists
to change the field's value.
In all top vg read functions only LCK_VG_READ/WRITE can be used.
All other vg lock definitions are low-level backend machinery.
Moreover, LCK_WRITE cannot be tested through bitmask.
This patch fixes these mistakes.
For _recover_vg() we do not need lock_flags, it can be only
two of above and we always upgrading to LCK_VG_WRITE lock there.
(N.B. that code is racy)
There is no functional change in code (despite wrong masking
it produces correct bits:-)
One shiny day we should use libblkid here. But now using LUKS is
very common together with LVM and pvcreate destroys LUKS completely.
So for user's convenience, try to detect LUKS signature and allow abort.
This is not only undocumented but is is also in violation with --help
documentation.
Using --yes without --force is useful in pvcreate when it detects
old signature.
pvcreate detects MD and swap signature.
The logic hidden there is not only documented but it is also
user unfriendly. Who invented this logic should run pvcreate
on its own critical MD device to see why;-)
This patch
- creates one function instead of duplication code
- asks if user want to overwrite signature
- allows aborting (!)
(Please note that writing LVM signatute without wiping old
is wrong, it confuses blkid, MD will not work anyway and
swap and LUKS is broken too.)
We can't rely on the fact that udev should prepare the node with right major
and minor number to trigger the module autoloading. We have to take into
account that the node could be missing or it could exist with improper
major and minor number assigned (e.g. from previous kernel versions in
an environment with static nodes and without udev). Make any corrections
if needed!
The lvm repair issues I believe are the superficial symptoms of this
bug - there are worse issues that are not as clearly seen. From my
inline comments:
* If the mirror was successfully recovered, we want to always
* force every machine to write to all devices - otherwise,
* corruption will occur. Here's how:
* Node1 suffers a failure and marks a region out-of-sync
* Node2 attempts a write, gets by is_remote_recovering,
* and queries the sync status of the region - finding
* it out-of-sync.
* Node2 thinks the write should be a nosync write, but it
* hasn't suffered the drive failure that Node1 has yet.
* It then issues a generic_make_request directly to
* the primary image only - which is exactly the device
* that has suffered the failure.
* Node2 suffers a lost write - which completely bypasses the
* mirror layer because it had gone through generic_m_r.
* The file system will likely explode at this point due to
* I/O errors. If it wasn't the primary that failed, it is
* easily possible in this case to issue writes to just one
* of the remaining images - also leaving the mirror inconsistent.
*
* We let in_sync() return 1 in a cluster regardless of what is
* in the bitmap once recovery has successfully completed on a
* mirror. This ensures the mirroring code will continue to
* attempt to write to all mirror images. The worst that can
* happen for reads is that additional read attempts may be
* taken.
Ignore snapshots when performing mirror recovery beneath an origin.
Pass LCK_ORIGIN_ONLY flag around cluster.
Add suspend_lv_origin and resume_lv_origin using LCK_ORIGIN_ONLY.
DM devices were not handled properly on nodes in a cluster that were not
where the splitmirrors command was issued. This was happening because
suspend_lv/resume_lv were being used in a place where activate_lv should
have been used.
When the suspend/resume are issued on (effectively) new LVs, their
'resource' (UUID) is not located in the lv_hash. Thus, both operations
turn into no-ops. You can see this from the output of clvmd from one
of the remote nodes:
<snip>
do_suspend_lv, lock not already held
<snip>
do_resume_lv, lock not already held
'activate_lv' enjoins the other nodes in the cluster to process the lock
and activate the new LV. clvmd output from remote node as follows:
do_lock_lv: resource 'zMseY7CBuO3Ty09vXlplPAHzD0Y0CovjrTdv0R1VcwggMwPdYhutHErRcwm5Nd2S', cmd = 0x19 LCK_LV_ACTIVATE (READ|LV|NONBLOCK), flags = 0x84 (DMEVENTD_MONITOR ), memlock = 1
sync_lock: 'zMseY7CBuO3Ty09vXlplPAHzD0Y0CovjrTdv0R1VcwggMwPdYhutHErRcwm5Nd2S' mode:1 flags=1
sync_lock: returning lkid 27b0001
Signed-off-by: Jonathan Brassow <jbrassow@redhat.com>
Reviewed-by: Petr Rockai <prockai@redhat.com>
This can happen with older rules (without support for synthesized events)
that are still part of initrd while using new udev rules in the system itself.
The consequence was that new udev rules incorrectly assumed that not having
DM_UDEV_PRIMARY_SOURCE_FLAG set always means the uevent is synthesized and
inappropriate (device is still not properly activated) and so it should be
ignored. However, initrd is not updated automatically while updating the
libdevmapper/udev rules in the system and so we end up with the rules not
detecting and setting crucial parts in the initrd environment and the rules
in the system that rely on the information that should have been stored in
udev db (which is incorrect in this configuration, of course).
The overall consequence is that the update of libdevmapper/lvm2 without
regenerating the initrd could end up with a boot failure! Ignoring the event
means removing any existing symlinks in /dev!
To fix this, increase udev rules version to make a difference. So from now on,
mark rules without proper support for synthesized events as
DM_UDEV_RULES_VSN="1" and 2 (or higher) if that support is included.
We still need to detect this one! We're not so strict with CHANGE events as
with the ADD events while applying filters in the rules so this one would
pass and it would process the rules prematurely (because it appears *before*
the actual CHANGE event used when resuming a DM device while setting read-only
state at the same time).
clvmd daemon itself does the right thing when invoked as non-root, by
returning 4.
The patch removes the use daemon function from
/etc/rc.d/init.d/functions that´s unnecessary and has th bad habit to
mask the return codes from the real daemon.
Add a simple and generic check to see if clvmd is executed by root or not.
Our stop/reload/restart paths in the init script are complex and not all
the tools involved in the process are guaranteed to return 4 if executed
by non-root against a process that´s running as root (for example kill
-TERM will return -1 and parsing the output to catch the error is
suboptimal at best).
https://bugzilla.redhat.com/show_bug.cgi?id=553381
The new standard in the storage industry is to default alignment of data
areas to 1MB. fdisk, parted, and mdadm have all been updated to this
default.
Update LVM to align the PV's data area start (pe_start) to 1MB. This
provides a more useful default than the previous default of 64K (which
generally ended up being a 192K pe_start once the first metadata area
was created).
Before this patch:
# pvs -o name,vg_mda_size,pe_start
PV VMdaSize 1st PE
/dev/sdd 188.00k 192.00k
After this patch:
# pvs -o name,vg_mda_size,pe_start
PV VMdaSize 1st PE
/dev/sdd 1020.00k 1.00m
The heuristic for setting the default alignment for LVM data areas is:
- If the default value (1MB) is a multiple of the detected alignment
then just use the default.
- Otherwise, use the detected value.
In practice this means we'll almost always use 1MB -- that is unless:
- the alignment was explicitly specified with --dataalignment
- or MD's full stripe width, or the {minimum,optimal}_io_size exceeds
1MB
- or the specified/detected value is not a power-of-2
Introduce --norestorefile to allow user to override the new requirement.
This can also be overridden with "devices/require_restorefile_with_uuid"
in lvm.conf -- however the default is 1.
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
We can already detect MD devices internally. But when using MD partitions,
these have "block extended major" (blkext) assigned (259). Blkext major
is also used in general, so we need to check whether the original device
is an MD device actually.
An incorrect fix on July 13, 2010 for an annoyance has caused a regression.
The offending check-in was part of the 2.02.71 release of LVM. That
check-in caused any PVs specified on the command line to be ignored when
performing a mirror split.
This patch reverses the aforementioned check-in (solving the regressions)
and posits a new solution to the list reversal problem. The original
problem was that we would always take the lowest mimage LVs from a mirror
when performing a split, but what we really want is to take the highest
mimage LVs. This patch accomplishes that by working through the list in
reverse order - choosing the higher numbered mimages first. (This also
reduces the amount of processing necessary.)
Signed-off-by: Jonathan Brassow <jbrassow@redhat.com>
Reviewed-by: Takahiro Yasui <takahiro.yasui@hds.com>
corruption bug in cmirror. 'dm_bit' is only ever used as a boolean operation
within LVM, but it can return a range of values. If the bit is set, a power of
2 is returned. If the bit is unset, 0 is returned.
'log_test_bit' (a function in the cluster mirror log daemon code) has switched
to using the dm bit operations in rhel6. There are two places in the daemon
code where 'log_test_bit' is not used merely as a boolean, but rather the
return value is used as the return value for the log functions 'is_clean' and
'in_sync' - having assumed that 'dm_bit' was returning 0 or 1 only.
One place the 'in_sync' function is utilized is in 'dm_rh_get_state' - a
function that informs the mirroring code how to treat I/O and which devices to
read/write from. 'dm_rh_get_state' was checking if the return value of
'in_sync' was 1 to determine if the region was DM_RH_CLEAN. Since 'dm_bit'
(and by extension 'log_test_bit' and 'in_sync') was returning powers of 2,
DM_RH_CLEAN was rarely being reported as it should have been. Thinking the
region was out-of-sync, the mirroring code would write only to the primary
device. When the primary device was failed, all of those writes were lost -
leaving the entire mirror corrupted.
udev_sync feature requires semaphores (part of System V IPC) to be configured
in kernel (CONFIG_SYSVIPC). Check whether it is supported and if not, give
a warning message and disable udev synchronisation code automatically to
avoid any further error states and associated problems.
One should use the kernel with System V IPC support enabled or libdevmapper
with udev_sync feature disabled.
all but one mirror leg.
<patch header>
To handle a double failure of a mirrored log, Jon's two patches are
commited, however, lvconvert command can't still handle an error
when mirror leg and mirrored log got failure at the same time.
[Patch]: Handle both devices of a mirrored log failing (bug 607347)
posted: https://www.redhat.com/archives/lvm-devel/2010-July/msg00009.html
commit: https://www.redhat.com/archives/lvm-devel/2010-July/msg00027.html
[Patch]: Handle both devices of a mirrored log failing (bug 607347) -
additional fix
posted: https://www.redhat.com/archives/lvm-devel/2010-July/msg00093.html
commit: https://www.redhat.com/archives/lvm-devel/2010-July/msg00101.html
In the second patch, the target type of mirrored log is replaced with
error target when remove_log is set to 1, but this procedure should be
also used in other cases such as the number of mirror leg is 1. This
patch relocates the procedure to the main path.
In addition, I added following three changes.
- Removed tmp_orphan_lvs handling procedure
It seems that _delete_lv() can handle detached_log_lv properly
without adding mirror legs in mirrored log to tmp_orphan_lvs.
Therefore, I removed the procedure.
- Removed vg_write()/vg_commit()
Metadata is saved by vg_write()/vg_commit() just after detached_log_lv
is handled. Therefore, I removed vg_write()/vg_commit().
- With Jon's second patch, we think that we don't have to call
remove_mirror_log() in _lv_update_mirrored_log() because will be
handled remove_mirror_images() in _lvconvert_mirrors_repaire().
</patch header>
Signed-off-by: Takahiro Yasui <takahiro.yasui@hds.com>
Reviewed-by: Petr Rockai <prockai@redhat.com>
Signed-off-by: Jonathan Brassow <jbrassow@redhat.com>
The cluster log daemon (cmirrord) is not multi-threaded and
can handle only one request at a time. When a log is stacked
on top of a mirror (which itself contains a 'core' log), it
creates a situation that cannot be solved without threading.
When the top level mirror issues a "resume", the log daemon
attempts to read from the log device to retrieve the log
state. However, the log is a mirror which, before issuing
the read, attempts to determine the 'sync' status of the
region of the mirror which is to be read. This sync status
request cannot be completed by the daemon because it is
blocked on a read I/O to the very mirror requesting the
sync status.
With mirror_log_fault_policy of 'remove' and mirror_image_fault_policy
of 'allocate', the log type of the mirror volume is converted from
'disk' or 'mirrored' to 'core' when all mirror legs but one in a mirror
volume broke.
Keep new_log_count as a number of valid log devices by using log_count
variable for a temporary usage in the first phase of error recovery
in _lvconvert_mirrors_repair().
Signed-off-by: Takahiro Yasui <takahiro.yasui@hds.com>
Reviewed-by: Petr Rockai <prockai@redhat.com>
There was missing "revert" call in _create_and_load_v4 fn while the preparation
for table load ends up with failure in create/load/resume sequence. Otherwise
we could end up with a device being created, but not table-loaded nor resumed.
Even though the table is not loaded and the device is not resumed at this
stage, we still need to synchronize with udev when calling the revert
"remove" ioctl - there's still a remove uevent generated! The "revert"
code does exactly that.
CMIRRORD_PIDFILE is not defined. This makes the build fail.
Therefore, we need to conditionalize the check for cmirrord
based on if CMIRRORD_PIDFILE is defined.
mirrors, we must also check that the log daemon (cmirrord) is running.
The log module can be auto-loaded, but the daemon cannot be
"auto-started". Failing to check for the daemon produces cryptic
messages that customers have a hard time deciphering. (The system
messages do report that the log daemon is not running, but people
don't seem to find this message easily.)
Here are examples of what is printed when the module is available,
but the log daemon has not been started.
[root@bp-01 LVM2]# lvcreate -m1 -l1 -n lv vg
Shared cluster mirrors are not available.
[root@bp-01 LVM2]# lvcreate -m1 -l1 -n lv vg -v
Setting logging type to disk
Finding volume group "vg"
Archiving volume group "vg" metadata (seqno 3).
Creating logical volume lv
Executing: /sbin/modprobe dm-log-userspace
Cluster mirror log daemon is not running
Shared cluster mirrors are not available.
Creating volume group backup "/etc/lvm/backup/vg" (seqno 4).
When splitting off mirror images from a mirror, we always take
LVs from the end of a list. For example, if the mirror sub-devices
are lv_mimage_[012], we should select lv_mimage_2 if splitting off
one image. However, lv_mimage_0 was being selected instead.
The problem came from calling '_move_removable_mimages_to_end'
when it was unnecessary to do so. When the user /does/ specify
specific devices to be removed, this function properly moved the
appropriate LVs to the end of the list for extraction. However,
if the user /doesn't/ give any specific PVs, the function should
do nothing. '_move_removable_mimages_to_end' was keying off of
whether 'removable_pvs' was NULL or not and this value was
improperly being populated with the set of all available PVs.
This was causing '_move_removable_mimages_to_end' to completely
reverse the list, which in turn caused us to extract the
hithertofore front-of-the-list LVs.
An unhandled condition allowed the command to terminate
cleanly without a warning. Added a check for the
'--splitmirrors' argument to allow execution to the lower
level function that has the check to see if the user is
trying to split a linear device. You should now see a
message if you try to use --splitmirrors on a linear device.
The main problem with these bugs was that the newly split
off LV was not being suspended properly. This meant that
the memlock count was not being balanced, the DM devices
were not being renamed, and some DM devices which should
have been removed were not.
I've also renamed some of the variables and added comments
to make things clearer as to what is going on. (I can break
this patch in two if it means easier review.)
pvchange: Add --metadataignore description
vgchange: Fix minor formatting
pvcreate: Update metadataignore description to refer to pvchange
lvm.conf: Refer to pvcreate and pvchange for metadata options.
Switch dmeventd to use dm_create_lockfile and drop duplicate code.
Allow clvmd pidfile to be configurable.
Switch cmirrord and clvmd to use dm_create_lockfile.
This should bring less confusion when there are some settings left and
people just forgot about it and then they run into problems. These messages
should give them a hint of what's really going on.
even though there was no log. A simple run through the in-tree test
suite would have caught this. :(
- if (lv_is_mirrored(detached_log_lv) &&
+ if (detached_log_lv && lv_is_mirrored(detached_log_lv) &&
Also, made some cosmetic changes suggested by kabi after my last check-in
(e.g. s/return 0/return_0/ and adding an error message).
A previous check-in added logic to handle the case where both images
of a mirrored log failed. It solved the problem by simply removing
the log entirely - leaving the parent mirror with a 'core' log. This
worked for most cases. However, if there was a small delay between
the failures of the two mirrored log devices, the mirror would hang,
LVM would hang, and no additional LVM commands could be issued.
When the first leg of the log fails, it signals the need for repair.
Before 'lvconvert --repair' is run by dmeventd, the second leg fails.
'lvconvert' would see both devices as failed and try to remove the
log entirely. When it came time to suspend the parent mirror to
update the configuration, the suspend would hang because it couldn't
get any I/O through the mirrored log, which was plugged waiting for
corrective action. The solution is to replace the log with an error
target to clear any pending writes before removing it. This allows
the parent mirror to suspend and make the proper changes.
Pass metadataignore through PV creation / setup paths.
As a result of this cleanup, we can remove the unnecessary setting
of mda_ignore bits inside pvcreate_single(), after call to pv_create.
For now, just set metadataignore to '0' in some places. This is
equivalent to the prior functionality, although the 0 is given
by the caller not hardcoded in _mda_setup() call.
Signed-off-by: Dave Wysochanski <dwysocha@redhat.com>
Indent updates.
Use AC_HELP_STRING for help string.
Start help string with lower letter.
Add [] around some default values i.e. [TYPE=internal].
Skip some "" around shell assigment when not needed.
Fix typo --with-device-gid=UID string.
When using vgmetadatacopies value other than "umanaged" (0), prompt
the user if the usage of --metadataignore would change the value of
vgmetadatacopies. The main 2 cases are:
1) pvchange --metadataignore
2) vgextend --metadataignore
We leave the prompt check in the tools, and do not change anything
if the user says 'n'.
Examples:
vgextend --metadataignore y vgtest /dev/loop0
Setting metadataignore will override preferred number of copies of VG vgtest metadata.
Are you sure? [y/n]: y
No physical volume label read from /dev/loop0
Physical volume "/dev/loop0" successfully created
Volume group "vgtest" successfully extended
pvchange --metadataignore y /dev/loop3
Setting metadataignore on /dev/loop3 will override preferred number of copies of VG vgtest metadata.
Are you sure? [y/n]: y
WARNING: Changing preferred number of copies of VG vgtest metadata from 3 to 2
Physical volume "/dev/loop3" changed
1 physical volume changed / 0 physical volumes not changed
Signed-off-by: Dave Wysochanski <dwysocha@redhat.com>
Test the auto-repair capability when we fail committing to an mda
on a new pv adding to a vg. This test should fail until we fix
the auto-repair in this case.
For now, this is just a precaution. Normally, all the other (non-dm) rules
should check DM_UDEV_DISABLE_OTHER_RULES_FLAG and therefore avoid setting
any inotify watches as well. But let's make sure.
Support for final assignment of the "nowatch" rule (the use of ":=") will
appear in next udev release, v160. This should also work in previous udev
versions but the setting won't be sealed so any further OPTIONS="watch" will
always prevail there.
We may want to add more specific "nowatch" rules later if needed.
- If a PV contained empty mdas, the auto-recovery code was not kicking in.
- The 'inconsistent' state was getting lost when metadata was cached so
recovery didn't kick in. But leave the behaviour alone when using
precommitted metadata because of a warning in a confusing FIXME.
In my testing, pvs and vgs didn't repair inconsistent metadata like they
used to do. (How many other tools fail similarly now?)
And there should be no need to cache inconsistent metadata because it is
supposed to get repaired under the protection of a write lock immediately it is
discovered.
This code is in need of a redesign based on first principles.
I still see bugs in this code and this commit is risky.
Rather than attempting to remove all the images of a mirrored
log volume via remove_mirror_images, simply remove the log
if all its devices have failed.
Taka was the first to report that there is still an outstanding
issue with handling this case. I've managed to reproduce it
only very rarely, and am still working on identifying the problem.
Failing to handle the problem rarely is better than not handling
the scenario at all, so I'm checking this in.
Moreover, in current mirror handling, when it calls activate
on removed but suspended detached log this counter drops below zero
and confuses debug log.
When a mirror is being downconverted in a cluster, a series of suspends and
resumes is executed.
With the change to using UUIDs in dev_manager instead of names, the behaviour
has changed with regards to including an _mlog in the deptree of a logical
volume. In the old (pre-UUID-enabled) code, the _mlog would appear in a deptree
of any volume purely based on a name match: a linear volume foo would include
foo_mlog in its dependencies if that happened to exist. This behaviour was
fixed and the mlog is now only included for mirrors.
By a coincidence, this mlog bug had been hiding a different bug in clvmd. When
a mirror is being dismantled (and converted to a linear volume), it is first
suspended as a whole, then later resumed in parts. Nevertheless, the overall
memlock balance is maintained in this operation. The problem kicks in, because
even though the mirror log was suspended as part of the mirror, when the
dismantled mirror is resumed again, it is no longer a mirror and therefore the
mirror log stays suspended. This would not be a problem in itself, since
_delete_lv (from metadata/mirror.c) is called on it subsequently, which does an
activate/deactivate cycle and removes the LV. The activate/deactivate cycle
correctly prompts clvmd to resume the device: however, in doing this, it will
issue an unpaired resume operation (the suspend that caused the mirror log to
be suspended is paired with resuming the dismantled mirror later). We have
concluded that the path in clvmd should never affect memlock_count, since there
should never be an unmatched explicit suspend preceding this resume.
Allow metadataignore flag to be passed in to pvcreate.
Ideally, more refactoring of the mda allocation / initialization
is warranted, but for now, we just add another parameter to 'add_mda'
to take an existing mda ignored flag. We need to do this or pv_write
loses the state of the mda 'ignored' flag before copying and writing
to disk.
Print device name when setting or clearing metadata ignore bit.
Example:
label/label.c:160 /dev/loop2: lvm2 label detected
cache/lvmcache.c:1136 lvmcache: /dev/loop2: now in VG #orphans_lvm2 (#orphans_lvm2)
metadata/metadata.c:4142 Setting mda ignored flag for metadata_locn /dev/loop2.
format_text/text_label.c:318 Skipping mda with ignored flag on device /dev/loop2 at offset 4096
Logging isn't ideal, especially for mda_set_ignore. Ideally we'd
like to display the device name and offset in this case but this
requires a bit more work and a per-format 'mda_description' function
pointer definition (we don't have access to mda_context in
metadata.c).
In preparation to call this from both pvcreate as well as pvchange,
move the guts of metadataignore into a library function.
Signed-off-by: Dave Wysochanski <dwysocha@redhat.com>
There's an intermittent failure with vgcfgbackup that seems to have been
introduced with the metadataignore / vgmetadatacopies patchset.
Intermittent failures are often the result of uninitialized data,
so this patch calls zalloc in a few places it might matter.
Update example.conf to describe vgmetadatacopies. Provide an
explanation for the '0' ("unmanaged") value.
Signed-off-by: Dave Wysochanski <dwysocha@redhat.com>
Allowing an 'all' and 'unmanaged' value is more intuitive, and
provides a simple way for users to get back to original LVM behavior
of metadata written to all PVs in the volume group.
If the user requests "--vgmetadatacopies unmanaged", this instructs
LVM not to manage the ignore bits to achieve a specific number of
metadata copies in the volume group. The user is free to use
"pvchange --metadataignore" to control the mdas on a per-PV basis.
If the user requests "--vgmetadatacopies all", this instructs LVM
to do 2 things: 1) clear all ignore bits, and 2) set the "unmanaged"
policy going forward.
Internally, we use the special MAX_UINT32 value to indicate 'all'.
This 'just' works since it's the largest value possible for the
field and so all 'ignore' bits on all mdas in the VG will get
cleared inside _vg_metadata_balance(). However, after we've
called the _vg_metadata_balance function, we check for the special
'all' value, and if set, we write the "unmanaged" value into the
metadata. As such, the 'all' value is never written to disk.
Signed-off-by: Dave Wysochanski <dwysocha@redhat.com>
Now that we have both --pvmetadatacopies and --vgmetadatacopies,
we need to make sure --metadatacopies gets interpreted correctly.
For pv commands, --metadatacopies should imply --pvmetadatacopies,
and for vg commands, --vgmetadatacopies.
Note: this will change the behavior of vgcreate with --metadatacopies
to be a synonym for --vgmetadatacopies. Previously, --metadatacopies
would apply to any PVs given with vgcreate that needed an implicit
pvcreate. As a result, one small change is needed to one of the nightly
tests - t-vgcreate-usage.
Signed-off-by: Dave Wysochanski <dwysocha@redhat.com>
When vgmerge is called we move the mdas from the source to the
destination. With metadata balancing we now have another mda
list, fid->metadata_areas_ignored, so move the mdas on this list
as well.
This patch should not matter as the code is written today. However
we include it for completeness in the case that _vgmerge_single()
is refactored and/or moved into a library function.
Signed-off-by: Dave Wysochanski <dwysocha@redhat.com>
The check in vg_split_mdas will trigger an error if the 'from' vg
list is empty. However, this might be ok in some instances now
that we have ignored mdas. Relax this check so an error is triggered
only in the case where there's truly no more mdas in the 'from'
vg.
One example of where this makes a difference is with vgreduce.
If we try to vgreduce a PV with un-ignored mdas, this should trigger
the balancing function to un-ignore mdas on another PV in the VG.
However, we don't get to vg_write() before we fail because this
list size check fails, and we see an error message indicating:
"Cannot remove final metadata area ..."
Another example is with vgsplit into a new VG, where the PVs
being moved contain all ignored mdas. We must move the mdas on
fid->metadata_areas_ignored from 'vg_from' to 'vg_to'.
Signed-off-by: Dave Wysochanski <dwysocha@redhat.com>
The vgextend path calls add_pv_to_vg(). Inside add_pv_to_vg(),
we must ensure we pass the correct mdas list into pv_setup(), as
copies of mdas are placed on the vg->fid list. If we don't place
the mdas on the correct vg->fid list, the various counts may be
incorrect and the metadata balance algorithm will not work when
called from vg_write() path.
Signed-off-by: Dave Wysochanski <dwysocha@redhat.com>
Allow parsing of --vgmetadatacopies for vgcreate. Accept
--metadatacopies as a synonym for --vgmetadatacopies.
Signed-off-by: Dave Wysochanski <dwysocha@redhat.com>
When a user explicitly sets a new mda ignore value for a PV, we
should update vg_mda_copies accordingly. When the VG is written
out, the user would not want the new ignore state to get lost as
a result of the vg_mda_copies value and logic in the vg_write
path.
Signed-off-by: Dave Wysochanski <dwysocha@redhat.com>
Compare the value of the newly added vg_mda_copies field
(--vgmetadatacopies parameter) with the current count of
in-use mdas and ignoring or unignoring mdas as necessary to
get to the target count. Also, as a safety check before
returning, ensure we have at least one mda enabled.
Signed-off-by: Dave Wysochanski <dwysocha@redhat.com>
Update logic in vgchange to handle --vgmetadatacopies, allow
--metadatacopies as a synonym to --vgmetadatacopies,
and add these parameters to args.h and commands.h
Forbit both --vgmetadatacopies and --metadatacopies as only
one allowed.
Signed-off-by: Dave Wysochanski <dwysocha@redhat.com>
This patch adds the ability to read/write the vg->mda_copies values
from/to the vg metadata.
If we read the VG metadata and this field does not exist, we set
mda_copies to the default value of 0. Later in the code, we use
this special '0' value to indicate a disable of metadata balancing.
This should preserve existing LVM behavior and ensure metadata balancing
can be turned off should the need arise.
Signed-off-by: Dave Wysochanski <dwysocha@redhat.com>
This patch adds the get and partially implemented set function.
The 'set' function should probably ignore or un-ignore metadata areas
based on new values.
Signed-off-by: Dave Wysochanski <dwysocha@redhat.com>
Add a field to struct volume_group to later implement metadata
balancing:
- mda_copies: target # of non-ignored mdas in the VG; default 0 (do
not control pv 'ignore mdas' bit.
This patch just adds the parameter to the structures with the default
values but does not modify any commands. Should be no functional change.
Signed-off-by: Dave Wysochanski <dwysocha@redhat.com>
This patch adds a vgmetadatacopies parameter for metadata balancing.
This parameter provides a simple way for users to create a policy for
placing metadata on PVs automatically by LVM. The behavior is implemented
inside LVM by managing the 'ignore' mda bits. We chose the name
'vgmetadatacopies' as this is a natural extension to the existing parameter
'pvmetadatacopies' / 'metadatacopies' in pvcreate.
This is a first step at VG parameter based metadata balancing. Most users
will probably want to state that they want a certain number of PVs to contain
metadata, and they may be less concerned about a specific number of metadata
copies in the volume group. However, for default values (pvmetadatacopies
is 1 by default), the number of metadatacopies in the volume group, and the
number of PVs with metadata are the same. In the future we could add
vgmetadatacopiespvs to define more specifically the number of pvs in the
VG that contain metadata, but for now we start with this parameter.
Another possible future extension would be to define a specific pv tag
to mark the set of PVs that should be used for metadata balancing. This
tag based approach could be used in conjunction with 'vgmetadatacopies'.
Signed-off-by: Dave Wysochanski <dwysocha@redhat.com>
Arrange mdas so mdas that are to be ignored come first. This is an
optimization that ensures consistency on disk for the longest period of time.
This was noted by agk in review of the v4 patchset of pvchange-based mda
balance.
Note the following example for an explanation of the background:
Assume the initial state on disk is as follows:
PV0 (v1, non-ignored)
PV1 (v1, non-ignored)
PV2 (v1, non-ignored)
PV3 (v1, non-ignored)
If we did not sort the list, we would have a commit sequence something like
this:
PV0 (v2, non-ignored)
PV1 (v2, ignored)
PV2 (v2, ignored)
PV3 (v2, non-ignored)
After the commit of PV0's mdas, we'd have an on-disk state like this:
PV0 (v2, non-ignored)
PV1 (v1, non-ignored)
PV2 (v1, non-ignored)
PV3 (v1, non-ignored)
This is an inconsistent state of the disk. If the machine fails, the next
time it was brought back up, the auto-correct mechanism in vg_read would
update the metadata on PV1-PV3. However, if possible we try to avoid
inconsistent on-disk states. Clearly, because we did not sort, we have
a greater chance of on-disk inconsistency - from the time the commit of
PV0 is complete until the time PV3 is complete.
We could improve the amount of time the on-disk state is consistent by simply
sorting the commit order as follows:
PV1 (v2, ignored)
PV2 (v2, ignored)
PV0 (v2, non-ignored)
PV3 (v2, non-ignored)
Thus, after the first PV is committed (in this case PV1), on-disk we would
have:
PV0 (v1, non-ignored)
PV1 (v2, ignored)
PV2 (v1, non-ignored)
PV3 (v1, non-ignored)
This is clearly a consistent state. PV1 will be read but the mda will be
ignored. All other PVs contain v1 metadata, and no auto-correct will be
required. In fact, if we commit all PVs with ignored mdas first, we'll
only have an inconsistent state when we start writing non-ignored PVs,
and thus the chances we'll get an inconsistent state on disk is much
less with the sorted method.
Signed-off-by: Dave Wysochanski <dwysocha@redhat.com>
When we are constructing the vg, we may need to adjust the list of
metadata_areas if there are ignored mdas. At label read time, we
do not read the metadata of ignored mdas, and as a result, they do
not get placed on vg->fid->metadata_areas inside _text_create_text_instance
since lvmcache does not have these areas attached to vginfo->infos.
However, when we're checking the pvids inside _vg_read, after having
read another metadata area from another PV, we do have the opportunity
to update the metadata_area and metadata_areas_ignored lists based
on the read metadata_area. We need accurate mda lists for the reporting
functions that count the ignored mdas, as well as general correctness
of mda balancing.
Signed-off-by: Dave Wysochanski <dwysocha@redhat.com>
With the addition of ignored mdas, we replace all checks for an empty
mda list with a new function to look for either an empty mda list or
ignored mdas.
Signed-off-by: Dave Wysochanski <dwysocha@redhat.com>
Add a helper function to consolidate checking for an empty mdas list
or ignored mdas. Ignored mdas should behave almost identically to
an empty mda list - the metadata areas should not be read or written
to. This function will make it easier to implement metadata balancing
and easier to track pvs with an empty mda list or ignored mdas.
Signed-off-by: Dave Wysochanski <dwysocha@redhat.com>
We implement ignore of an mda at label_read time by checking for
the ignore bit, and then skipping the reading of the vgname and
other information in the metadata. This will have an effect similar
to a PV found with no mdas. Thus, it will look like an orphan in the
cache until we scan the rest of the system and find a PV with
metadata, and the mda will not be on the vg->fid->metadata_areas
list so no read/writes will be done to the metadata area.
Signed-off-by: Dave Wysochanski <dwysocha@redhat.com>
This patch just modifies pvchange to call the underlying ignore
functions for mdas. Ensure special cases do not reflect changes
in metadata (PVs with 0 mdas, setting ignored when already ignored,
clearing ignored when not ignored).
Signed-off-by: Dave Wysochanski <dwysocha@redhat.com>
Define a new pvs field, pv_mda_used_count, and a new vgs field,
vg_mda_used_count to match the existing pv_mda_count and vg_mda_count.
These new fields count the number of mdas that have the 'ignored' bit
clear (they are in use on the PV / VG). Also define various supporting
functions to implement the counting as well as setting the ignored
flag and determining if an mda is ignored. These high level functions
call into the lower level location independent mda ignore functions
defined by earlier patches.
Note that counting ignored mdas in a vg requires traversing both lists
and checking for the ignored bit on the mda. The count of 'ignored'
mdas then is defined by having the bit set, not by which list the mda
is on. The list does determine whether LVM actually does read/write to
the mda, though we must count the bits in order to return accurate numbers
for the various counts. Also, pv_mda_set_ignored must search both vg
lists for ignored mda. If the state changes and needs to be committed
to disk, the ignored mda will be on the non-ignored list.
Note also in pv_mda_set_ignored(), we must properly manage the mda lists.
If we change the ignored state of an mda, we must change any mdas on
vg->fid->metadata_areas that correspond to this pv. Also, we may
need to allocate a copy of the mda, as is done when fid->metadata_areas
is populated from _vg_read(), if we are un-ignoring an ignored mda.
Signed-off-by: Dave Wysochanski <dwysocha@redhat.com>
Add a second mda list, metadata_areas_ignored to fid, and a couple
functions, fid_add_mda() and fid_add_mdas() to help manage the lists.
These functions are needed to properly count the ignored mdas and
manage the lists attached to the 'fid' and ultimately the 'vg'.
Ensure metadata_areas_ignored is initialized in other formats, even
if the list is never used.
Signed-off-by: Dave Wysochanski <dwysocha@redhat.com>
Because of the way mdas are handled internally, where a PV in a VG
has mdas on both info->mdas and vg->fid->metadata_areas list, we
need a location independent copy constructor for struct
metadata_area. Break up the existing format-text specific copy
constructor into a format independent piece and a format dependent
piece.
This function is necessary to properly implement pv_set_mda_ignored().
Signed-off-by: Dave Wysochanski <dwysocha@redhat.com>
Reviewed-by: Alasdair G Kergon <agk@redhat.com>
A metadata_area is defined independent of the location. One downside
is that there is no obvious mapping from a pv to an mda. For a PV in
a VG, we need a way to start with a PV and end up with an MDA, if we
are to manage mdas starting with a device/pv. This function provides
us a way to go down the list of PVs on a VG, and identify which ones
match a particular PV.
I'm not entirely happy with this approach, but it does fit into the
existing structures in a reasonable way.
An alternative solution might be to refactor the VG - PV interface such
that mdas are a list tied to a PV. However, this seemed a bit tricky since
a PV does not come into existence until after the list of mdas is
constructed (see _vg_read() - we create a 'fid' and attach mdas to it,
then we go through them and attach pvs).
Signed-off-by: Dave Wysochanski <dwysocha@redhat.com>
Reviewed-by: Alasdair G Kergon <agk@redhat.com>
We'd like to pass in mda_header to vgname_from_mda(). In order to
do this, we need to call raw_read_mda_header() from text_label.c,
_text_read(), which gets called from the label_read() path, and
peers into the metadata and update vginfo cache. We should check
the disable bit here, and if set, not peer into the vg metadata,
thus reducing the I/O to disk.
In the process, move vgname_from_mda() to layout.h, since the fn
only gets called from format_text code, and we need the mda_header
definition from the private layout.h.
Signed-off-by: Dave Wysochanski <dwysocha@redhat.com>
This refactoring moves the device open/close up one level to the caller of
_vg_read_raw_area(). Should be no functional change and facilitate future
changes related to metadata balancing.
Signed-off-by: Dave Wysochanski <dwysocha@redhat.com>
First we add a 'flags' field to the location independent
metadata_area structure, and a MDA_IGNORE flag. The
mda_is_ignored and mda_set_ignored functions are added to
manage the flag. Adding the flag and functions gives a
library interface to ignore metadata areas independent of
the underlying location (disk, file, etc). The location
specific read/write functions must then handle the specifics
of what this flag means to the location.
Signed-off-by: Dave Wysochanski <dwysocha@redhat.com>
Reviewed-by: Alasdair G Kergon <agk@redhat.com>
Adding a flag to the 'rlocn' structure in the mda header of the
text format allows us to flip a bit to ignore an area on disk that
stores the metadata via the text format specific mda_header.
This patch defines the flag and access functions to manage the flag.
Other patches will manage the ignore on a format-independent basis,
by using a flag in the metadata_area structure.
Signed-off-by: Dave Wysochanski <dwysocha@redhat.com>
Future patches will make use of a specific flag in the on-disk 'raw_locn'
structure to enable/disable metadata areas, and facilitate metadata
balancing.
Note that 'filler' is always set to '0' (see add_mda() - memset),
so use of this area as a non-zero flags field is a safe way to
provide future code features.
Signed-off-by: Dave Wysochanski <dwysocha@redhat.com>
The same region size is used for both mirror volume and mirrored
log volume, but when the physical extent size is bigger than region size,
the size of mirror leg for mirrored log is smaller than the region size
and lvcreate command fails.
This patch adjusts a region size of mirrored log to a smaller value of
region size or physical extent size.
[This patch ensures that the region_size of the mirrored log does not
exceed the size of the mirrored log itself, which would violate the
kernel constraint: (region_size <= ti->len).]
Signed-off-by: Takahiro Yasui <takahiro.yasui@hds.com>
Signed-off-by: Jonathan Brassow <jbrassow@redhat.com>
Preload libc.mo file for localized lvm before taking memory lock - this way
we prevent disk access for some error paths in libdm, that prints localized
errno messages while they are still in memory locked state.
the failure of a device that contained both a image of
a mirror and an image of the mirrored log. The order
of the handling of those faults was important (and
wrong), this patch corrects that.
Patch-From: Takahiro Yasui <tyasui@redhat.com>
We can use DM_UDEV_PRIMARY_SOURCE_FLAG to identify the spurious events
and use it as an indication that the device has already been activated before
(and hence we can find this property in udev database).
WARNING: This change requires udev startup script to preserve udev database
from initrd. All the information stored there during activation of devices
is important for the initial "udevadm trigger --action=add" call that is
used in udev startup script. If not done this way, udev startup script needs
to define DM_UDEV_PRIMARY_SOURCE_FLAG=1 property for any ADD events it uses.
- s/Active clustred VG/clustered VG/ (only LV can be active)
- print only active LVs (not all) in status command
(In the lvdisplay form /dev/vg/lv.)
For now, still use awk (already used in clustered_vgs).
https://bugzilla.redhat.com/show_bug.cgi?id=598495
converting from 2-way to 3-way mirror (collapse_mirrored_lv)
was calling '_remove_mirror_images' with the 'remove_log'
parameter set. When the code was put in to fix 599898 to
honor log parameters during conversion, this argument was
suddenly being honored. Thus, when someone would convert from
a 2-way to 3-way mirror, the log would get removed.
'collapse_mirrored_lv' should not be calling '_remove_mirror_images'
with 'remove_log' set.
to 3-way mirror. When conversion operations are performed on
these types of mirrors, log options can be confused/ignored.
In the case of a converting 3-way mirror, we have a top-level
2-way corelog mirror whose legs are 1) a 2-way disk-log mirror
and 2) a linear device. If we wish to convert this 3-way mirror
to a 2-way mirror, the linear device is removed and the extra
top layer is eliminated. If we also wished to convert the disk
log to a core log in the same step, ambiguity creeps in. It is
somewhat obvious what the user wants - a 2-way mirror with a
corelog. However, looking at the top level mirror before
compression, it seems that the mirror already has a core log.
This is why the operation seemed to fail.
This patch simply re-evaluates what mirrored_seg points to after
a compression and then considers the log argument.
This is a fix for bug 599898.
Because execve stops the command loop,
we never receive response (only socket close) for clvmd -S,
so waiting for response here makes no sense.
But if the calling process (clvmd -S) exits too early, connection
is closed from client side, clvmd takes this as an error and
never run restart code.
Ugly hack(TM).
When using clustered mirrors, we need device nodes to be created during
processing of device tree, not at its end like we normally do (we need to
access the nodes in cmirror prematurely). Therefore we use a new flag called
"immediate_dev_node" stored in deptree's load_properties struct to instruct the
device tree processing code to immediately synchronize with udev and flush all
stacked node operations so the nodes are prepared for use.
For now, the immediate_dev_node is used for clustered mirrors during
processing the dm_tree_preload_children code only. We can add more later if
needed.
linux/kdev_t.h even though it wasn't needed. Strangely, it seems
to be causing problems on various architectures (i686) in the
function daemons/cmirrord/functions.c:disk_status_info()->sprintf.
I'm not sure why this is a problem since none of the macros in
kdev_t.h are used in that code, but it certainly doesn't hurt to
pull an unnecessary header and it seems to fix the problem.
Code is mixing up internal DLM and LVM definitions of lock
modes and flags.
OpenAIS and singlenode locking do not depend on DLM but
code currently cannot be compiled without libdlm.h!
LCK_* flags is LVM abstraction, used through all the code.
Only low-level backend (clvmd-cman etc) should use DLM definitions,
also this code should do all needed conversions.
Because there are two DLM flags used in generic code
(NOQUEUE, CONVERT) we define it similar way like lock modes.
(So all needed binary-compatible flags are on one place in locking.h)
(Further code cleaning still needed, though:-)
Introduce lvm_exec_prefix with resolved exec_prefix.
(using same ac_default_prefix as for CLVMD_PATH)
Use lvm_exec_prefix instead of dmeventd_prefix (fixes missing ac_default_prefix)
Note: This patch is rather hot-fix as currently generate code
does not create correct code for make exec_prefix=
- allocate environment dynamically (still missing some limit?)
- try to recover, if destroy failed (do not destroy lvm here) and free memory
- check strdup() return codes
- report failure to log
- do not print NULL in exclusive lock loop
A kernel patch is on its way for 2.6.35 adding support for dm-mod module
autoload. Udev v155 and higher is able to read static node information given
in modules.devname (extracted by depmod before) and will create such nodes
at its start. The first access to such node will load the module automatically
(directly in kernel) before the actual read/write operation is processed.
Activate only the first replicator-dev LV, that activates all other
related LVs from Replicator. In case of error during this activation,
it will not retry again for other heads (less confusing error log).
length(array) is specific to GNU awk and doesn't work in mawk.
Use a return value of "split" function to indicate array size, this is
supported in both gawk and mawk.
This patch fixes the following errors during "make install" when mawk is
installed as a default awk.
mawk: scripts/relpath.awk: line 25: illegal reference to array from
mawk: scripts/relpath.awk: line 25: illegal reference to array to
mawk: scripts/relpath.awk: line 27: illegal reference to array from
mawk: scripts/relpath.awk: line 32: illegal reference to array to
Signed-off-by: Mikulas Patocka <mpatocka@redhat.com>
Adding function _add_partial_replicator_to_dtree() to create
partial tree for Replicator target.
Using dm_tree_node_set_presuspend_node() for Replicator.
As for _process_one_vg() we need similar retry loop for
process_each_lv_in_vg(). This patch retries to process
failed LVs with reopened VGs.
Patch does not add any extra repeated invocations if there is not
found any missing VG during LV processing.
Patch modifes behavior of _process_one_vg().
In the first pass vg_read() collectis for replicator sorted list of
additional VGs during lock_vol().
If any other VG is needed by the replicator and it is not yet opened
then next iteration loop is taken with all collected VGs.
Flag vg->cmd_missing_vgs detects missing VGs.
Introduce struct cmd_vg to store information about needed
volume group name, vgid, flags and the pointer to opened VG.
Keep VGs list in alphabetical order for locking order.
Introduce functions:
cmd_vg_add() add new cmd_vg entry.
cmd_vg_lookup() search cmd_vgs for vg_name.
cmd_vg_read() open VGs in cmd_vgs list.
cmd_vg_release() close VGs in reversed order.
Add pointer to linked list of opened VGs. List temporarily keeps
the information about needed or locked and opened VGs for replicator target.
Also add cmd_missing_vgs flag information for quick check and
also for possible continuos process_each_lv() usage where we need
to detect whether failure has been caused by missing VG or
some other reason.
Adding configure.in support for Replicators.
Adding basic lib lvm support for Replicators.
Adding flags REPLICATOR and REPLICATOR_LOG.
Adding segments SEG_REPLICATOR and SEG_REPLICATOR_DEV.
Adding basic methods for handling replicator metadata.
For deactivation of Replicator check in advance that all heads
have open_count == 0. For this presuspend_node is used as all
head nodes are linking this control node.
Introducing dm_tree_node_set_presuspend_node() for presuspending child
node (i.e. replicator control target) before deactivation of parent node
(i.e. replicator-dev target).
This patch presents no functional change to current dtree - only
replicator target currently sets presuspend node for dev nodes.
Patch adds failed_lvnames to the list of parameters for process_each_lv_in_vg().
If the list is not NULL it will be filled with LV names of failing LVs
during function execution.
Application could later reiterate only on failed LVs.
lvm2app forces applications to start with a volume group name,
open the volume group, then operate on individual pvs. In some
cases the application may want to start with a device name rather
than the volume group name. Today, if an application wants to
do this, it must iterate through all the volume groups to find
the volume group that the specific device is attached to.
These new interfaces allow the application to avoid such overhead.
Bump the lvm2app version number to 3.
Earlier patches added some infrastructure to lookup a vgname from
a pvname. We now can cleanup some of the pvchange and other code
by requiring callers that want to modify some pv property:
1) lookup the vgname by the pvname
2) use the vgname to obtain a vg handle
3) get the pv handle from the vg handle
This should work going forward and be a much cleaner interface,
as we move away from pvs as standalone objects.
Some commands start with a pvname, but we'd like to force users to
start with a vg handle to obtain a pv handle. Our best option seems
to be providing a way to look up the vgname from the pvname, and then
require them to use vg_read/vg_open.
In addition to the pvname lookup function, this patch also provides a
lookup by pvid. The lookup by pvid can be used in conjunction with
lvmcache_get_pvids to process all pvs in the system.
The pvid find function first calls lvmcache_vgname_from_pvid, which may
cause the label to be read if it is not in the cache. If the vgname is
returned is an orphan, we then check to see if there are metadata areas,
and if not, we scan every PV on the system by calling scan_vgs_for_pvs().
In most cases we should not need to do this, and by using the info->mdas
count, we avoid calling pv_read() as prior code did. So this patch is a
bit cleaner and should allow us to refactor more of the pv code.
Signed-off-by: Dave Wysochanski <dwysocha@redhat.com>
are active mirrors or snapshots.
We don't have the mechanisms in place to change the device-mapper
tables for those targets that have behavioral differences between
cluster and single machine instances. Allowing users to change
the attribute but not changing the target's behavior can lead to
data corruption.
The following bugs are fixed/avoided by this patch:
235123 - vgchange -c [ny] do not change target types when necessary
289331 - RFE: switching from cluster domain to local domain needs to deactivate volume somehow
289541 - when changing from local to cluster, volumes can not appear to be deactivated
This should avoid various races between dmeventd on multiple nodes
in cluster where one node already repairing device and another
run full scan and locks the device.
the device cache file is dumped both in vgscan and clvmd process.
Unfortunately, clvmd calls lvmcache_label_scan,
it properly destroys persistent filter, but during
persistent_filter_dump it merges old cache content back!
This causes that change in filters is not properly propagated
into device cache after vgscan on cluster.
(Only new devices are added.)
https://bugzilla.redhat.com/show_bug.cgi?id=591861
to a file using log/file, with log/overwrite set and dump this file in
STACKTRACE. The overall effect is that only the command that ran last before
the failure has been triggered will get its debug output logged. This is
similar to how we treat coredumps.
lvconvert testing is now in its own test, t-lvconvert-mirror-basic ... it
doesn't do anything fancy but it does run lvconvert through a lot of
combinations.
I have also merged the remaining t-mirror-lvconvert tests into
t-lvconvert-mirror and abolished the former. The latter will be split again
later into more thematic divisions. (The previous split was rather arbitrary,
may I even say random...)
Use Requires.private: instead of Libs.private:
Use UDEV_PC and SELINUX_PC for Require.private:
It looks like usage of Requires.private is prefered from Libs.private.
However pkg-config documentation is really poor here. But here is
short outcome:
There is a difference in Libs.private: and Requires.private: where
we specify libselinux instead of -lselinux -lsepol.
We leave resolving of query like 'pkg-config --libs --static devmapper'
on taking proper selinux and udev libs to their .pc files instead of
hardcoding them into our .pc file which is might give incorrect answer.
- i.e. dependency of libselinux package might change and we may return
wrong list of linked libraries.
http://bugs.freedesktop.org/show_bug.cgi?id=4738http://err.no/personal/blog/tech/2008-03-25-18-07_pkg-config,_sonames_and_Requires.private
A shortcut for --ignorelockingfailure, --ignoremonitoring, --poll n options
and LVM_SUPPRESS_LOCKING_FAILURE_MESSAGES environment variable used all at
once in initialisation scripts (e.g. rc.sysinit or initrd).
Target install_dm_plugin installs files to libdir/device-mapper.
Target install_lvm2_plugin installs files to libdir/lvm2.
Both targets creates relative links to libdir to keep the code
compatible with current dlopen handling.
Once we will be able to read plugins from subdir, links
could be removed.
(with table provided).
This remove ioctl generates udev events like any other hence it needs to be
synchronized properly as well. Also, add dm task type in debug log when
setting a cookie (for better debugging).
fails, the test will carry on but will issue a warning. The harness detects
such warnings from tests and marks tests that passed with warnings with a
special status.
We can use it even in read-only environment where a try to initialise
file-based locking fails (not to mention other processing related with
lvm2 init). Simply, we want to output the version only, nothing else.
And this should always work.
is active and while it is in-active.
+for i in $(seq 0 4); do
+ for j in $(seq 0 4); do
+ for k in core disk mirrored; do
+ for l in core disk mirrored; do
The testing code still needs some improvement. I'd like to add
the ability to test specifying the PVs to be added/removed during
a convert. It will also be important to test partial PV
specification during down converts (i.e. request to remove more
mirror images than we have provided PVs for).
This rule appeared in udev v152 and it helps us to support spurious events
where we didn't have any flags set (events originated in udevadm trigger
or the watch rule). These flags are important to direct the rule application.
Now, with the help of this rule, we can regenerate old udev db content.
To implement this correctly, we need to flag all proper DM udev events with
DM_UDEV_PRIMARY_SOURCE_FLAG. That happens automatically for all ioctls
generating events originated in libdevmapper.
being able to remove more images from a mirror than the
number of PVs directly specified for removal.
The effort to fix bug 581611 corrected a bug that was unnoticed
at the time. The loop in _remove_mirror_images that looks over
the specified PVs was allowing devices that were previously
counted and moved to the end of the list to be double-counted.
This resulted in the number of devices needed for removal always
being satisfied - even if the user did not specify enough PVs
for removal to satisfy the request. When 581611 was fixed, this
double-counting no longer took place and the result was to remove
only the minimum of the number of PVs specified or the number
that was asked to be removed.
By simply always setting 'new_area_count' (as used to be done
only in the else statement), we return to the previous behavior.
Indeed, this is exactly what the double-counting was allowing
to happen before the fix of 581611.
Allow lv_remove_with_dependencies() to know the top-level LV that was
requested to be removed (otherwise it recurses and we lose context).
A merging snapshot cannot be removed directly but the associated origin
can be. Disallow removal of a merging snapshot unless the associated
origin is also being removed.
(process_each_lv_in_vg) provides it.
Also, removing this check from _lvs_single now allows displaying hidden
LVs that are specifically named on the command line.
There's no need for foreign udev rules to touch LVM reserved devices
(snapshot, pvmove, _mlog, _mimage, _vorigin) even if they happen to
be visible. The same applies for /dev/disk content - no need to create
any content for these devices (and so no need to run any "blkid" etc.).
This also prevents setting any inotify "watch" from udev rules on such
devices that is a source of race conditions (the rules need to honor
DM_UDEV_DISABLE_OTHER_RULES_FLAG for this to work though).
1) Test that the primary mirror image cannot be removed while
the mirror set is sync'ing.
2) Test that you cannot start a second mirror up-convert while
one is already in progress.
The trouble is that if the sync/conversion finishes before the
tests occur, the tests will fail by why of success where there
should have been failure. This means the sync/conversion must
happen very quickly, but this is possible because the test
mirrors we are creating are so small.
In order to decrease the likelyhood of these test failing (or
more correctly, failing to test the right thing), I've increase
the size of the mirrors. It will still be remotely possible that
the tests will fail (by way of failing to test the right thing).
If this continues to happen, more involved mechanisms will need
to be put in place. (Perhaps these will still be created, but
this change should be a remedy until that time.)
These two old-test/regex utils are usable for testing output of regex
processing at core level. For getting them usable configure need to create
Makefile. This is currently disable by default.
Reintroduce split teardown (teardown() calls teardown_devs()) because
t-topology-support.sh only needs the teardown_devs() subset of the full
teardown() between each iteration of the topology tests -- in particular
the $TESTDIR must not get removed between each topology test iteration.
prepare_loop() must return if prepare_scsi_debug_dev() already
established $LOOP.
Also fix (and simplify) the unsafe scsi-debug device discovery in
prepare_scsi_debug_dev().
Ensure we can create devices for use in tests before running the real
tests. This may in various cases, and may involve machine configuration
rather than a failure in some specific test.
For example, if a test is run with LVM_TEST_DIR=/tmp, selinux is enabled,
and the default security context is set to "<<none>>" for /tmp, all the
tests will fail, unable to create devices, since dmsetup will fail, a
result of machpathcon() returning an error code.
Signed-off-by: Dave Wysochanski <dwysocha@redhat.com>
This version number change reflects the memory handling change
for string-based pv/vg/lv string based attributes.
In addition, when adding support for tags, I forgot to increase
the version number.
Signed-off-by: Dave Wysochanski <dwysocha@redhat.com>
Everywhere else in the API the caller can rely on lvm2app taking care of
memory allocation and free, so make the 'name' and 'uuid' properties of a
vg/lv/pv use the vg handle to allocate memory.
Signed-off-by: Dave Wysochanski <dwysocha@redhat.com>
We should write metadata into next position in the ring buffer while calling
vgrename and vgcfgrestore. At this code level (_vg_write_raw), we were not able
to determine if this is a rename or not. If yes, then accompanying VG structure
passed here has a new name set, not the old one.
When looking for a location where to put metadata next, we were given a NULL
value because of failed VG name comparison (in _find_vg_rlocn) between the
name in existing metadata and metadata we're just about to write.
This resets the position in the ring buffer, overwriting any existing metadata
(and also incorrectly updates the cache to "orphan" afterwards).
This patch just adds old_name item in struct volume_group that we can check and use
if necessary and detect renames at lower layers as well.
The same applies for vgcfgrestore, but here we're using a special value of
old_name, an empty string, to disable the check with existing metadata totally.
Internally, we used DM names instead of UUIDs while processing event
handlers. This caused problems while trying to vgrename a VG with active LVs
where the names are being changed and so the devices were not found then.
The patch also contains a little bit of refactoring, moving "build_dlid" code
found in dev_manager.c to "build_dm_uuid", now in lvm-string.c (so we have
build_dm_uuid and build_dm_name at one place).
lvm2app needs a link back to the vg in order to use the vg handle for
memory allocations as well as other things. This patch adds the field
to struct physical_volume, and sets pv->vg when reading a vg from disk or
extending a vg by using the helper function previously added,
add_pvl_to_vgs(). Moves and renames are handled with separate code
inside move_pv() and vgmerge(). Add pv->vg check to vg_validate().
A NULL value in pv->vg signifies membership in the orphan VG.
Note though in the case of pv_read() on a device with metadatacopies == 0,
more devices may need to be read for an authoritative answer.
Signed-off-by: Dave Wysochanski <dwysocha@redhat.com>
Now that we have library functions to add/delete a pv from the vg->pvs
list, call them from everywhere.
Signed-off-by: Dave Wysochanski <dwysocha@redhat.com>
Add a delete function to manage the vg->pvs list.
NOTE: It may be possible to do further cleanup to these add/del functions
by passing a 'pv' as input instead of 'pv_list'. The pv_list is used for
functions which do allocations (lvcreate) while other places in the code
just manage a list of 'pv' (e.g. import functions, vgextend, etc).
Signed-off-by: Dave Wysochanski <dwysocha@redhat.com>
Move the increment of vg->pv_count next to the place where we add to
vg->pvs. It looks safe to do this since the only caller of import_pool_vg()
calls import_pool_pvs() immediately afterward, and there is no way
import_pool_vg() can fail (always returns 1). However, if there's a
memory allocation failure inside import_pool_pvs(), we will end up with
a different count in vg->pv_count that with the original code. In any
case, vg->pv_count should be as close to dm_list_size(&vg->pvs) as
possible, as is the case everywhere else in the code.
Signed-off-by: Dave Wysochanski <dwysocha@redhat.com>
The dm_list * parameter is unnecessary since we are passing in 'vg'
and the only caller of import_pool_pvs() passes '&vg->pvs' in the
dm_list * parameter. Just use vg->pvs directly in the function.
Signed-off-by: Dave Wysochanski <dwysocha@redhat.com>
{core|disk|mirrored}. However, under the options section they are
described as {core|disk} - even though the 'mirrored' argument is
described.
s/'{core|disk}'/'{core|disk|mirrored}'/
Fix unwanted modification of $(top_builddir)/make.tmpl.
Using dependency rules to install rules for udev.
There is minor problem, with concurent usage of builddir
and srcdir could lead to missuse of 10-dm.rules which
could be found in VPATH from different builddir.
However current solution uses intermediate target so
the generated 10-dm.rules exists only for short period of time
during make install execution.
Patch is inspired by Debian's extra patch.
- removes OWNER & GROUP make vars they are parts of INSTALL command.
- adds INSTALL_PROGRAM for executable, uses $(INSTALL)
- adds INSTALL_DATA for non-executable data, uses ($INSTALL)
- adds INSTALL_WDATA for writable non-executable data, uses ($INSTALL)
- adds configure option --enable-write_install - to support
installatin of writable files used by distribution
- replaces usage of ifeq @LIB_SUFFIX@ with $(LIB_SUFFIX)
- installs .a files from static builds without executable flag
- installs .a files to $(usrlibdir) instead of $(libdir)
- installs all static binaries to $(staticdir)
- create .so links for devel package in $(usrlibdir) instead of
$(libdir)
- makes .so and .so.LIB_VERSION files within builddir
- removes VERSIONED_SHLIB and created versioned LIB_SHARED automagicaly
- install LIB_SHARED via install_lib_shared target
- install plugins via install_lib_shared_plugin target
- prints whole 'install' command during installation instead of less
informative "Installing $(something) $(somewhere)"
- install multiple man pages with one INSTALL command
- use DISTCLEAN_TARGETS instead of creating multiple distclean targets
Usage of VPATH makes troubles when used within $(builddir).
Not only source files are being found through VPATH,
but targets as well. (make --debug=v)
Thus if user builds the code in $(srcdir) and also in some $(builddir)
he gets mangled results as some generated files (i.e. .export.sym)
are 'reused' from $(srcdir) instead of $(builddir).
This patch switches to use vpath were we could explicitly name
suffixes that should be looked via vpath - we must take care,
we do not generate files with these suffixes:
.c, .in, .po, .exported_symbols
A user specifying duplicate paths on the cmdline of vgcreate will
get a message similar to the following:
vgcreate vgtest2 /dev/loop3 /dev/loop5
Found duplicate PV jk1lXsKzwyOKlXq6bhaFFKMQQ06oPgu8: using /dev/loop5 not /dev/loop3
Found duplicate PV jk1lXsKzwyOKlXq6bhaFFKMQQ06oPgu8: using /dev/loop3 not /dev/loop5
Internal error: Duplicate PV id jk1lXs-Kzwy-OKlX-q6bh-aFFK-MQQ0-6oPgu8 detected for /dev/loop3 in vgtest2.
This is caught by vg_validate(), but it would be good to find
this condition earlier in the vgcreate code. add_pv_to_vg()
currently checks by pvname, but does not look for duplcate pvids.
This patch adds the check for duplicate pvids and results in new
error output as follows:
vgcreate vgtest2 /dev/loop3 /dev/loop5
Found duplicate PV jk1lXsKzwyOKlXq6bhaFFKMQQ06oPgu8: using /dev/loop5 not /dev/loop3
Found duplicate PV jk1lXsKzwyOKlXq6bhaFFKMQQ06oPgu8: using /dev/loop3 not /dev/loop5
Physical volume '/dev/loop5 (jk1lXs-Kzwy-OKlX-q6bh-aFFK-MQQ0-6oPgu8)' listed more than once.
Unable to add physical volume '/dev/loop5' to volume group 'vgtest2'.
Signed-off-by: Dave Wysochanski <dwysocha@redhat.com>
When moving parts of striped LVs, pvmove wouldn't care about leaving you with
two stripes on the same disk. Now --alloc anywhere is needed for that.
(Tried and gave up on two alternative approaches before the one committed here.)
Small refactor of main places in the code where a pv is added to a
vg into a small function which adds the pv to the list and updates
the vg counts.
Signed-off-by: Dave Wysochanski <dwysocha@redhat.com>
Refactor adding to the vg->pvs list and incrementing the count, which
will allow further refactoring. Should be no functional change.
Signed-off-by: Dave Wysochanski <dwysocha@redhat.com>
Simple refactor to mov code that updates the vg extent counts from a
single pv's counts close to the code that adds a pv to vg->pvs and
updates vg->pv_count. No functional change.
Signed-off-by: Dave Wysochanski <dwysocha@redhat.com>
In add_pv_to_vg(), we should only add the pv to vg->pvs after all
internal checks have passed. The check for vg->extent_count exeeding
maximum was after we added the pv to the list, so this function could
return a state of vg->pvs that did not reflect other parameters such
as vg->pv_count.
Signed-off-by: Dave Wysochanski <dwysocha@redhat.com>
to check for presence of this module and avoid using --frames
option for genhtml in this case.
Fix arg list for AC_PATH_PROG for lcov and genhtml.
(detecting empty LCOV and GENHTML string in Makefiles).
Patch fixes generation of coverage files for dmeventd and adds support for clvmd.
Path names are stripped, so the the html looks better.
Frames 'previews' is enabled for generated pages.
Using top_srcdir was wrong here - though we still can't easily use builddir.
Requiers using shell variables before execution of binaries build outside
of srcdir.
Because we have now strong rule for lock ordering:
- VG locks must be taken in alphabetical order
- ORPHAN locks must be the last
vgs_locked() is now not needed.
This fixes problem with orphan locking, e.g.
vgremove VG1 | vgremove VG2
lock(VG1) | lock(VG2)
lock(ORPHAN) | lock(ORPHAN) -> fail, non-blocking
https://bugzilla.redhat.com/show_bug.cgi?id=578413
(More similar places in code.)
Physical segments were still allocated from global
command context mempool.
This leads to very high memory usage when
activating large VG (vgchange).
(Memory usage was about 2G when >3000LVs).
Fix it by properly using vg->vgmem private pool,
so all the memory is released early.
New memory pool parameter is needed here for pv_split_segment
function.
Also fix the same problem in some minor allocations
(vg description, lv segment split).
In addition to previous patch, we really do not need
to search for segment which was just allocated in
split request.
Make pv_split_segment function return newly allocated
(split) segment also.
(So after this patch, there is only one user
of slow find_peg_by_pe).
The function find_peg_by_pe is incredibly inefficient
for Pvs with many segments.
In shiny future there should be binary (or interval) tree
instead of sorted linked list (volunteers?).
Anyway, for now, we can use dirty trick here to optimise this case:
- Allocations are usually applied from the beginning
of PV (we have no alloocation policy which allocates areas
"backwards")
- The only user of find_peg_by_pe is pv_split_segment()
call. In *most* cases it need to split *last* PV segment.
So if we search sorted pv segment list backwards, we
hit the requested segment immediatelly.
This patch applies this tiny change.
(and saves >30% of processing time when >3000LVs segments are on one PV!)
To discourage using this inefficient function from other code,
it is moved to pv_manip.c and used static for now:-)
vg_validate call is an adept to optimisation, it is very
ineeficient and slow.
Anyway, we should call it only before writing data to disk.
The call in lvmcache was just temporary validation,
we realy do not need to revalidate cached metadata
every time.
(Actually, I added that there just to prove that cache works
properly and forgot to remove it.)
Patch removes it from lvmcache completely, this can hit only
internal bug in export function (and this bug must
be detected in any vg_write call anyway before).
The _read_vg uses already hash for PVs to optimise
reading of large VGs and avoiding repeated PV list traversing.
Use the same aproach to speed up parsing VG with many LVs.
If dmeventd runs with -d flag, it doesn't fork into backgroud.
The command kill(getppid(), SIGTERM) attempts to kill the parent dmeventd
process, however, if there is no parent, it kills whatever process spawned
dmeventd. In case of debugging with gdb, the parent is gdb, thus
kill(getppid(), SIGTERM) kills the debugger.
Signed-off-by: Mikulas Patocka <mpatocka@redhat.com>
we are running the repair manually. If we don't ignore, then dmeventd
and the manually run repair can collide. (We should still get clean
results in such a case, but it makes it harder to validate the test
results.)
Code moves initilization of stats values to _memlock_maps().
For dmeventd we need to use mlockall() - so avoid reading config value
and go with _use_mlockall code path.
Patch assumes dmeventd uses C locales!
Patch needs the call or memlock_inc_daemon() before memlock_inc()
(which is our common use case).
Some minor code cleanup patch for _un/_lock_mem_if_needed().
If the error path of _register_for_event() calls _free_thread_status()
_lib_put() call is missing.
To make thing simpler move this _lib_put() into common error path code.
As the header file <sys/mman.h> was not included in dmeventd.c
thus missed definition of MCL_CURRENT so this patch only makes
it obvious we were not locking memory here.
This patch has no functional change.
Later part of this patch set handles mlockall() via memlock_inc_daemon().
clvmd does not propagate DMEVENTD_MONITOR_IGNORE.
Update get_activation_monitoring_mode() to check if the VG that the
LV is being activated in is clustered. If so, skip it.
Any get_activation_monitoring_mode() error will cause the associated LV
(or VG) to be skipped during activation. Both vgchange_single() and
lvchange_single(), which call get_activation_monitoring_mode(), are
called by their respective process_each_..() method.
in clvmd, dmevend, man, tests.
Don't include dependency files for clow and cscope.out targets
Improve dependency tracking for dmeventd and liblvm2cmd sources.
to obtain sources. Create make.tmpl target for
simplier generation of cflow files with the help of
CFLOW_LIST, CFLOW_LIST_TARGET, CFLOW_TARGET.
Still cflow usage is not perfect.
Move daemons/ and lib/ subtargets to their Makefiles so we don't get
double cleanup error during execution of distclean target.
Instead of duplicating clean target inside distclean target,
just use it as a subtarget and avoid add duplicating code.
This check-in enables the 'mirrored' log type. It can be specified
by using the '--mirrorlog' option as follows:
#> lvcreate -m1 --mirrorlog mirrored -L 5G -n lv vg
I've also included a couple updates to the testsuite. These updates
include tests for the new log type, and some fixes to some of the
*lvconvert* tests.
clvmd's do_lock_lv() already properly controls dmeventd monitoring based
on LCK_DMEVENTD_MONITOR_MODE in lock_flags -- though one small fix was
needed for this to work: _lock_for_cluster() must treat
dmeventd_monitor_mode()'s return as a tri-state value.
Also cleanup do_lock_lv() to:
- explicitly init_dmeventd_monitor() based on LCK_DMEVENTD_MONITOR_MODE
- no longer reset init_dmeventd_monitor() to default at the end of
do_lock_lv() -- it is unnecessary
This is the next preparatory step towards better --alloc anywhere
support and is not intended to break anything that currently works so
please report any problems - segfaults, bogus data in the new debug
messages, or if the code now chooses bizarre allocation layouts.
. Add "monitoring" option to "activation" section of lvm.conf
. Have clvmd consult the lvm.conf "activation/monitoring" too.
. Introduce toollib.c:get_activation_monitoring_mode().
. Error out when both --monitor and --ignoremonitoring are provided.
. Add --monitor and --ignoremonitoring support to lvcreate. Update
lvcreate man page accordingly.
. Clarify that '--monitor' controls the start and stop of monitoring in
the {vg,lv}change man pages.
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
This prevents some confusion when libudev was not found so udev_sync was disabled
automatically. Configure was successful though giving only a tiny warning.
Also, if "dmsetup udevcreatecookie" is used, never return 0x000000 as a result if
udev is not running and keep the output blank.
We need to know whether we should wait for any uevent or not when
using udev_sync. A kernel patch was posted recently that changed the
way uevents are sent on dm device resume - it is sent only if the
device has been suspended before. There's also a new DM_UEVENT_GENERATED_FLAG
in the ioctl to notify userspace whether the event was generated.
If the uevent was not generated (e.g. the situation where the device is
*not* suspended and we call a resume), we just call dm_udev_complete
explicitly from within libdevmapper itself to prevent infinite waiting
while trying to synchronise with udev processing.
Prevent lvresize from being able to resize internal LVs: mirror legs
(*_mimage_*), mirror log (*_mlog), snapshot placeholder LVs (snapshot*)
and others. Resizing these would leads to unexpected metadata and
sometimes crashes (in case of growing snapshot*).
When we pv_read() a device that has an orphan vgname, we might need to scan
the system to be sure this is true. However, if the PV has mdas, there's
no way possible for it to have an orphan vgname unless it is a true orphan.
Some areas of the code were optimized to take advantage of this fact, while
others were not (we would still do the expensive scan if a device had mdas
but had an orphan VG).
This patch unifies the code so that every place we are operating on such
a PV, we skip the expensive scan if there are mdas.
Signed-off-by: Dave Wysochanski <dwysocha@redhat.com>
Acked-by: Petr Rockai <prockai@redhat.com>
Acked-by: Alasdair G Kergon <agk@redhat.com>
This option should be configurable, but for now
do not set it at all.
(lvm2app is used in udisks probers and there
cac cause several nasty races when trying to update
lvmcache during rescan.)
If user try to vgcreate or vgextend non-existent VG,
these messages appears:
# vgcreate xxx /dev/xxx
Internal error: Volume Group xxx was not unlocked
Device /dev/xxx not found (or ignored by filtering).
Unable to add physical volume '/dev/xxx' to volume group 'xxx'.
Internal error: Attempt to unlock unlocked VG xxx.
(the same with existing VG and non-existing PV & vgextend)
# vgextend vg_test /dev/xxx
...
It is caused because code tries to "refresh" cache if
md filter is switched on using cache destroy.
But we can change filters and rescan even without this
machinery now, just use refresh_filters
(and reset md filter afterwards).
(Patch also discovers cache alias bug in vgsplit test,
fix it by using better filter line.)
(VDSO on 32bit is VSyscall on 64bit)
It seems it could be locked on 64bit kernels running 32bit binaries,
but it makes troubles on real 32bit machines where mlock() returns
error when trying to lock such map area. (0xffffe000)
Behavior of mlockall() seems to be similar.
This patch adds a new implementation of locking function instead
of mlockall() that may lock way too much memory (>100MB).
New function instead uses mlock() system call and selectively locks
memory areas from /proc/self/maps trying to avoid locking areas
unused during lock-ed state.
Patch also adds struct cmd_context to all memlock() calls to have
access to configuration.
For backward compatibility functionality of mlockall()
is preserved with "activation/use_mlockall" flag.
As a simple check, locking and unlocking counts the amount of memory
and compares whether values are matching.
For static builds dependency for SELinux libs is not handled by 'ar'.
Till better solution is found, for static builds STATIC_LIBS is used.
Patch updates SELinux detection to use 3rd & 4th parameter for Success/Fail.
Also removes detection of pthread from this check as we know which
version of libdevmapper we are going to link with lvm after merge.
SELinux header check moved to the SELinux test code.
Create new substituted variable PTHREAD_LIBS and link this library
only with tools/libs which really needs it - i.e. dmeventd.
Check for libpthread only for builds with clvmd or dmeventd.
Remove variable LIB_PTHREAD
Modify linking of readline library. Create new substituted varible
READLINE_LIBS - readline library is linked ONLY with tools that really use
it - i.e. lvm. (Static lvm does not use readlin).
Previous behaviour put this library into the variable LIBS and thus
linked it with all created object files of lvm project (i.e. plugins...).
READLINE detection is simplified.
Termcap library is linked in only if readline library doesn't have its own
dependency (i.e. old distributions).
The kernel's blk_stack_limits() function may flag a device as
'misaligned'. If it does the alignment_offset will be -1.
Update set_pe_align_offset() to accommodate this corner case.
- increase timeout to 30 secs (on Chrissie request)
- source both cluster and clvmd for options (like all the other cluster
init scripts)
- add clustered_vgs and _lvs commodity fns
- move rh_status* fns at the top, so they can be reused
- heavily cleanup start and stop fns from redundant code and unnecessary
loops
- improve output from different operations
- make the init script lsb compliant
- don´t force kill of the daemon, send only a TERM signal and then wait
for it to exit
- Resolves rhbz#533247
lvm2 devices have always UUID set even if imported from lvm1 metadata.
Patch removes name argument from dev_manager_info call and converts
all activation related calls to use query by UUID.
Also it simplifies mknode call (which is the only user on mknodes parameter).
Fix add/remove tag function headers.
Fix a lot of little problems with doxygen comments.
Clarify the basic objects and their handles, and place functions with their
appropriate handles/objects.
All this cleanup moves automatic documentation of lvm2app much closer to being
useful as official documentation. In the future I will add some examples
and plan to build the examples as part of the unit tests.
Signed-off-by: Dave Wysochanski <dwysocha@redhat.com>
Add lvm2app functions to manage LV tags.
For lvm_lv_get_tags(), we return a list of tags, similar to other
functions that return lists. An empty list is returned if there
are no tags. NULL is returned if there is a problem obtaining
the list of tags.
Signed-off-by: Dave Wysochanski <dwysocha@redhat.com>
Add lvm2app functions to manage VG tags.
For lvm_vg_get_tags(), we return a list of tags, similar to other
functions that return lists. An empty list is returned if there
are no VG tags. NULL is returned if there is a problem obtaining
the list of tags.
Signed-off-by: Dave Wysochanski <dwysocha@redhat.com>
Add a supporting function to copy a list of internal tags to lvm2app list.
We need to put this here because of the lvm_str_list_t type which we export
in lvm2app.h. If we didn't export this type, we could put this in the
internal library and use struct str_list.
Signed-off-by: Dave Wysochanski <dwysocha@redhat.com>
We need to allocate memory for the tag and copy the tag value before we
add it to the list of tags. We could put this inside lvm2app since the
tools keep their memory around until vg_write/vg_commit is called, but
we put it inside the internal library to minimize code in lvm2app.
We need to copy the tag passed in by the caller to ensure the lifetime of
the memory until the {vg|lv} handle is released.
Signed-off-by: Dave Wysochanski <dwysocha@redhat.com>
Similar refactoring to vgchange - pull out common parts and put into
library function for reuse. Should be no functional change.
Signed-off-by: Dave Wysochanski <dwysocha@redhat.com>
Pull out common code to be called from tools as well as lvm2app.
Leave archive() at tool level so we can use from vgcreate
as well as vgchange. Should be no functional change.
- add stack macro in vgchange
Signed-off-by: Dave Wysochanski <dwysocha@redhat.com>
Add a merging snapshot to the deptree, using the "error" target, rather
than avoid adding it entirely. This allows proper cleanup of the -cow
device without having to rename the -cow to use the origin's name as a
prefix.
Move the preloading of the origin LV, after a merge, from
lv_remove_single() to vg_remove_snapshot(). Having vg_remove_snapshot()
preload the origin allows the -cow device to be released so that it can
be removed via deactivate_lv(). lv_remove_single()'s deactivate_lv()
reliably removes the -cow device because the associated snapshot LV,
that is to be removed when a snapshot-merge completes, is always added
to the deptree (and kernel -- via "error" target).
Now when the snapshot LV is removed both the -cow and -real devices
get removed using uuid rather than device name. This paves the way
for us to switch over to info-by-uuid queries.
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
There's a tiny period of time when the _mimage device is visible during
downconversion from mirror to linear. Since it is visible, we need to
create the symlinks, otherwise warning messages will be issued about udev
not creating those symlinks. We have to rely on udev flags completely.
- add DM_UDEV_DISABLE_LIBRARY_FALLBACK udev flag to rely on udev only
- export dm_udev_create_cookie function to create new cookies on demand
- add --udevcookie, udevcreatecookie and udevreleasecookie for dmsetup
(to support "udev transactions" where one cookie value can be used for
several dmsetup calls)
- don't use DM_UDEV_DISABLE_CHECKING env. var. anymore and set the state
automatically (based on udev and libdevmapper dev path comparison)
Internally we store sizes in sectors, but lvm2app exports sizes
in bytes. We could get fancier and allow units configuration but
this fix should do for now.
Fixes rhbz561422.
Signed-off-by: Dave Wysochanski <dwysocha@redhat.com>
We unfortunately don't yet _know_, in dev_manager_snapshot_percent(), if
a snapshot-merge target is active (activation is deferred if dev is
open); so we can't short-circuit origin devices based purely on existing
LVM LV attributes.
Set 'fail_if_percent_unsupported' in dev_manager_snapshot_percent() for
a merging origin LV, otherwise passing unsupported LV types to _percent
will lead to a default successful return with percent_range as
PERCENT_100.
For a merging origin, PERCENT_100 will result in a polldaemon that runs
infinitely (because completion is PERCENT_0).
When activating a merging origin it is valid, and expected, to not have
a node in the deptree for both the origin and its merging snapshot. The
_cached_info() caller is only concerned with whether a device is open.
If there isn't a node in the tree the associated device is definitely
not open.
This change was deferred to help ease the review of previous refactoring
related to using process_each_lv() for lvconvert's merge support. Not
that doing so _really_ helped but...
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Switch lvconvert's --merge code over to using process_each_lv(). Doing
so adds support for a single 'lvconvert --merge' to start merging
multiple LVs (which includes @tag expansion).
Add 'lvconvert --merge @tag' testing to test/t-snapshot-merge.sh
Adjust man/lvconvert.8.in to reflect these expanded capabilities.
The lvconvert.c implementation requires rereading the VG each iteration
of process_each_lv(). Otherwise a stale VG instance associated with
the LV passed to lvconvert_single_merge() would result in stale VG
metadata being written back out to disk. This overwrote new metadata
that was written when a previous snapshot LV finished merging (via
lvconvert_poll). This is only an issue when merging multiple LVs that
share the same VG (a single VG is typical for most LVM configurations on
system disks).
In the end this new support is very useful for performing a "system
rollback" that requires multiple snapshot LVs be merged to their
respective origin LV.
The yum-utils 'fs-snapshot' plugin tags all snapshot LVs that it creates
with a common 'snapshot_tag' that is unique to the yum transaction.
Rolling back a yum transaction, that created LVM snapshots with the tag
'yum_20100129133223', is as simple as:
lvconvert --merge @yum_20100129133223
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
refactoring.
Document the need to cleanup the "name" args passed around polldaemon,
lvconvert and pvmove. It is quite a mess.
Annotate the unused nature of the existing poll_fns->get_copy_vg
methods' 'uuid' arg.
depending on if the mirror has a 'core' or 'disk' log. When there
is a disk log, the new leg is added by stacking a new mirror on
top of the old (one leg is the old mirror and the other leg is the newly
added device). When the log is a 'core' log, the new leg is simply added
to the existing mirror and all the devices are re-synced.
The logic that handles collapsing the stacked 'disk' log mirror was
having the effect of causing 'core' logged mirrors to begin resync'ing
for a second time. I have used the 'CONVERTING' flag to indicate that
a mirror is converting by way of stacking. This is no longer set for
up-converting core logs. The final 'collapse' logic can safely be skipped
for 'core' log mirrors - getting rid of the second resync.
Signed-off-by: Jonathan Brassow <jbrassow@redhat.com>
where we should not expose internal VG names/uuids (the ones with "#" prefix )through the
interface. Otherwise, we could end up with library users opening internal VGs which will
initiate locking mechanism that won't be cleaned up properly.
"#orphans_{lvm1, lvm2, pool}" names are treated in a special way, they are truncated first
to "orphans" and this is used as a part of the lock name then (e.g. while calling lvm_vg_open()).
When library user calls lvm_vg_close(), the original name "orphans_{lvm1, lvm2, pool}"
is used directly and therefore no unlock occurs.
We should exclude internal VG names and uuids in the lists provided by lvmcache:
lvmcache_get_vgids() and lvmcache_get_vgnames().
Allow the number of logical extents to be expressed (for a snapshot) as
a percentage of the total space in the Origin Logical Volume with the
suffix %ORIGIN.
Update the relevant man pages accordingly. Eliminate inconsistencies
between the man pages and tools/commands.h
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
*_safe. This had the effect of segfaulting the log daemon when
converting a mirror from one log type to another.
Signed-off-by: Jonathan Brassow <jbrassow@redhat.com>
When activation of pvmove mirror fails on cluster, some nodes
still possibly succeeded in activation.
- Explicitly deactivate that mirror to be sure
- properly pair suspend/resume calls to not cause memory lock problems in clvmd
Code cannot simply call _finish_pvmove on cluster in this situation, because
changed LVs are suspended twice (causing memory inbalance) and also temporary
mirror is activated when it is not expected (and we know that it failed already).
Patch prepares special function which remove temporary mirror references from
metadata and then resumes changed LVs.
In dev_manager_info 0 means error and 1 info is returned,
not that device exists (that value is part of info struct).
Fix query by uuid only (no name) which returns 0 when device
does not exist.
Support "wait before testing" using '+' in pvmove and lvconvert
interval. Doing so overrides the new default of sleeping after checking
the LV's progress.
Sleeping before checking progress can lead to extraneous polldaemons
being left running. These polldaemons would have otherwise exited had
they checked before sleeping. Checking progress before sleeping helps
workaround the subtly unreliable nature of "finished" state checking
in _percent_run.
Update test/t-mirror-names.sh to use '+' when providing its lvconvert
interval.
more descriptive message if locking fails instead of
"Locking type -1 initialisation failed."
Use read-only locking instead of misleading ignorelocking option
in message.
Some files were not shown because too many files have changed in this diff
Show More
Reference in New Issue
Block a user
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
