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.
Change run time access to the command_name struct
cmd->cname instead of indirectly through
cmd->command->cname. This removes the two run time
fields from struct command.
All lvconvert functionality has been moved out of the
previous monolithic lvconvert code, except conversions
related to raid/mirror/striped/linear. This switches
that remaining code to be based on command defs, and
standard process_each_lv arg processing. This final
switch results in quite a bit of dead code that is
also removed.
This is a new explicit version of 'lvconvert LV'
which has been an obscure way of triggering polling
to be restarted on an LV that was previously converted.
Lift all the snapshot utilities (merge, split, combine)
out of the monolithic lvconvert implementation, using
the command definitions. The old code associated with
these commands is now unused and will be removed separately.
This lifts the lvconvert --repair and --replace commands
out of the monolithic lvconvert implementation. The
previous calls into repair/replace can no longer be
reached and will be removed in a separate commit.
The new check_single_lv() function is called prior to the
existing process_single_lv(). If the check function returns 0,
the LV will not be processed.
The check_single_lv function is meant to be a standard method
to validate the combination of specific command + specific LV,
and decide if the combination is allowed. The check_single
function can be used by anything that calls process_each_lv.
As commands are migrated to take advantage of command
definitions, each command definition gets its own entry
point which calls process_each for itself, passing a
pair of check_single/process_single functions which can
be specific to the narrowly defined command def.
. Define a prototype for every lvm command.
. Match every user command with one definition.
. Generate help text and man pages from them.
The new file command-lines.in defines a prototype for every
unique lvm command. A unique lvm command is a unique
combination of: command name + required option args +
required positional args. Each of these prototypes also
includes the optional option args and optional positional
args that the command will accept, a description, and a
unique string ID for the definition. Any valid command
will match one of the prototypes.
Here's an example of the lvresize command definitions from
command-lines.in, there are three unique lvresize commands:
lvresize --size SizeMB LV
OO: --alloc Alloc, --autobackup Bool, --force,
--nofsck, --nosync, --noudevsync, --reportformat String, --resizefs,
--stripes Number, --stripesize SizeKB, --poolmetadatasize SizeMB
OP: PV ...
ID: lvresize_by_size
DESC: Resize an LV by a specified size.
lvresize LV PV ...
OO: --alloc Alloc, --autobackup Bool, --force,
--nofsck, --nosync, --noudevsync,
--reportformat String, --resizefs, --stripes Number, --stripesize SizeKB
ID: lvresize_by_pv
DESC: Resize an LV by specified PV extents.
FLAGS: SECONDARY_SYNTAX
lvresize --poolmetadatasize SizeMB LV_thinpool
OO: --alloc Alloc, --autobackup Bool, --force,
--nofsck, --nosync, --noudevsync,
--reportformat String, --stripes Number, --stripesize SizeKB
OP: PV ...
ID: lvresize_pool_metadata_by_size
DESC: Resize a pool metadata SubLV by a specified size.
The three commands have separate definitions because they have
different required parameters. Required parameters are specified
on the first line of the definition. Optional options are
listed after OO, and optional positional args are listed after OP.
This data is used to generate corresponding command definition
structures for lvm in command-lines.h. usage/help output is also
auto generated, so it is always in sync with the definitions.
Every user-entered command is compared against the set of
command structures, and matched with one. An error is
reported if an entered command does not have the required
parameters for any definition. The closest match is printed
as a suggestion, and running lvresize --help will display
the usage for each possible lvresize command.
The prototype syntax used for help/man output includes
required --option and positional args on the first line,
and optional --option and positional args enclosed in [ ]
on subsequent lines.
command_name <required_opt_args> <required_pos_args>
[ <optional_opt_args> ]
[ <optional_pos_args> ]
Command definitions that are not to be advertised/suggested
have the flag SECONDARY_SYNTAX. These commands will not be
printed in the normal help output.
Man page prototypes are also generated from the same original
command definitions, and are always in sync with the code
and help text.
Very early in command execution, a matching command definition
is found. lvm then knows the operation being done, and that
the provided args conform to the definition. This will allow
lots of ad hoc checking/validation to be removed throughout
the code.
Each command definition can also be routed to a specific
function to implement it. The function is associated with
an enum value for the command definition (generated from
the ID string.) These per-command-definition implementation
functions have not yet been created, so all commands
currently fall back to the existing per-command-name
implementation functions.
Using per-command-definition functions will allow lots of
code to be removed which tries to figure out what the
command is meant to do. This is currently based on ad hoc
and complicated option analysis. When using the new
functions, what the command is doing is already known
from the associated command definition.
Kernel 4.10 (dm-crypt v1.15.0) and later supports loading device
tables with crypt segment having key in kernel keyring retention
service.
dmsetup hid key section of tables output. With this patch dmsetup
no longer hides key section if it detects kernel key description
instead of hex byte representation of key itself.
Commit cfb6ef654d introduced
support to change RAID region size.
Add:
- missing conditions to support any types to function with
it in lv_raid_convert(); temporary workaround used until
cli validation patches get merged
- tests requesting "-R " to lvconvert-raid-takeover.sh
involving a cleanup of the script
Related: rhbz1392947
Add:
- support to change region size of existing RaidLVs
(all RAID LV types but raid0/raid0_meta)
- lvconvert-raid-regionsize.sh with test variations
for different RAID types and region sizes
Resolves: rhbz1392947
Add a new update_filemap command to dmstats that allows a filemap
group to be updated:
# dmstats update_filemap --groupid 0 vm.img
/var/lib/libvirt/images/vm.img: Updated group ID 0 with 137 region(s).
This will update the set of regions mapped to the file to reflect
the current file system allocation.
Currently this needs to be run manually - a future update will add
support for monitoring file maps via a daemon, allowing them to be
automatically updated when the underlying file is modified.
This was missing piece in 77997c7673.
When merging origin is inactive (while driver is loaded) we
could already report merge in progress values as there is
no way to activate 'old state' now.
Solves: https://bugzilla.redhat.com/1280496
The only reasonable behaviour here is to error on
any number out of accepted range (i.e. now numbers
wrapping around with some hidden logic).
As this is plain bug there is no support for
backward compatibility since noone should
set numbers >UINT32_MAX and expect 0 or error
depending on how big number was used....
TODO: more fields might need to be converted.
Add simple function to wrap usage for only uint32 numbers.
Unlike 'int_arg' which accepts full range of 64bit number
this function will error on numbers out of this range:
<0, UINT32_MAX>
When there is 'merging' of an origin in progress, but metadata stil
do provide both origin and snapshot, we should show data from merged
snapshot. This is important mainly for thin case, where there was
a window, where i.e. 'lvs -o+device_id' would report information
about 'already gone' origin thin LV.
This race window is usually hard to trigger but can be ocasionally hit.
Usually shortly after activation, but before polling process manages
to update metadata after merge.
Before starting polling process, validate the merge has actually started
so there is not pointless invoke of lvmpolld.
This also fixes reported message from command, so user has
correct info whether merging has already started or
if it's delayed for next activation.
We can't keep 'display_lvname' for too long - it's using
ringbuffer and keeps limited number of names. So it's
safe only per few simple tests, but can't be used anymore
after some function calls..
(Fixes 00e641ef37)
Add this functionality to lvconvert:
'lvconvert --thin cachedLV --thinpool vg/poll'
Converts cachedLV to external origin (which will be read-only).
New thin volume is created in thinpool LV and it's using external
origin as source for unprovisioned chunks.
This conversion happens online (while volume is in use).
Thin LV remains fully writable.
Cached external origin no longer could be written so cache will be used
ONLY for read operations. For this limitation we require cache mode
to be writethrough (as writeback cannot write to read-only volumes).
When thinLV is later removed cached external origin is again
fully usable, just note, LV remain in 'read-only' mode.
When read-write is needed, 'lvchange -prw' has to be used.
Single external origin could be user by multiple thinLV in
multiple differen thin pool.
When --count=0 interval numbers are miscalculated:
Interval #18446744069414584325 time delta: 999920887ns
Interval #18446744069414584325 current err: -79113ns
End interval #18446744069414584325 duration: 999920887ns
This is because the command line argument is cast through the
uint32_t type, and fixed to UINT32_MAX:
_count = ((uint32_t)_int_args[COUNT_ARG]) ? : UINT32_MAX;
We also need to handle --count=0 specially when calculating the
interval number: since intervals count from #1, this must account
for the implicit "minus one" when converting from zero to the
UINT64_MAX value used (which is too large to store in _int_args).
The time management code mixes tests of the _timer_fd value with
code that should be timer agnostic: this causes problems for users
of the usleep() timer, since it cannot properly detect the start
of a new interval:
Beginning first interval
Interval #18446744069414584348 time delta: 1000000000ns
Interval #18446744069414584348 current err: 0ns
End interval #18446744069414584348 duration: 1000000000ns
Adjusted sample interval duration: 1000000000ns
[...]
Beginning first interval
Interval #18446744069414584349 time delta: 1000000000ns
Interval #18446744069414584349 current err: 0ns
End interval #18446744069414584349 duration: 1000000000ns
Adjusted sample interval duration: 1000000000ns
Separate these out, by defining a _timer_running() call that each
timer implements, and only define _timer_fd if we are compiling
with TIMERFD enabled.
Although the usleep() interval timer is not used if the Linux
TIMERFD interface is available it should still provide reasonably
good timing.
Instead of trying to estimate the error from the duration of the
last sleep, peg it to the start time of the program, and use the
value of ((start_time - now) % interval) to correct the current
interval duration.
This always pulls us back into sync at the end of each interval,
rather than relying on trying to incrementally adjust the time
duration at each interval start.
This greatly reduces drift when the usleep() clock is used.
Drop LV from passed API arg - it's always segment being checked.
Also use_layer is now in full control of lv_info_with_seg_status().
It decides which device needs to be checked to get 'the most info'.
TODO: future version should be able to expose status from
Start moving selection of status taken for a LV into a single place.
The logic for showing info & status has been spread over multiple
places and were doing too complex decision going agains each other.
Unify selection of status of origin & cow scanned device.
TODO: in future we want to grab status for LV and layered LV and have
both statuses present for display - i.e. when 'old snapshot'
of thinLV is takes and there is ongoing merge - at some moment
we are not capable to show all needed info.
pvscan --cache -aay was activating LVs in exported VGs
when it should not.
It appears that this was a regression from commit 9b640c3684
"pvscan: use process_each_vg for autoactivate".
(Automatic) repair may not be allowed during the initial sync of an upconverted
linear LV, because the data on the failing, primary leg hasn't been completely
synchronized to the N-1 other legs of the raid1 LV (replacing failed legs during
repair involves discontinuing access to any replaced legs data, thus preventing
data recovery on the primary leg e.g. via dd_rescue).
Even though repair would not cause data loss when adding legs to a fully synced
raid1 LV, we don't have information yet defining this state yet (e.g. a raid1
LV flag telling the fully synchronized status before any legs were added),
hence can't automatically decide to allow to repair.
If nonetheless a repair on a non-synced raid1 LVs is intended, the "--force"
option has to be provided.
Resolves: rhbz1311765
Validate kernel support for raid0/raid4 on given and
requested segtype before requesting conversions on them.
Because raid10 wasn't present in old RAID targets, add
the same validation to be prepared once we support them.
Check for dm-raid target version with non-standard raid4 mapping expecting the dedicated
parity device in the last rather than the first slot and prohibit to create, activate or
convert to such LVs from striped/raid0* or vice-versa in order to avoid data corruption.
Add related tests to lvconvert-raid-takeover.sh
Resolves: rhbz1388962
On conversions between striped/raid0* and raid4, the kernel expects
the dedicated raid4 parity SubLVs in the first segment area rather than
in the last it's been allocated to, thus the data mapping ain't proper.
Enhance lvconvert (lib/metadata/raid_manip.c) to shift the dedicated
parity SubLVs on conversions from striped/raid0* to raid4 and vice-versa.
In case of raid0_meta -> raid4 where the MD raid0 personality already has
stored RAID array device positions in the superblocks, the MetaLVs have to
be cleared so that the kernel doesn't fail validating the array positions
after lvm has shifted them up by one.
Add more tests to lvconvert-raid-takeover.sh including one to check for
mapping flaws by converting a created raid4 with filesystem -> striped
and fsck it.
Whilst on it:
- add missing direct striped -> raid4 conversion to the takeover array
to avoid an intermim conversion from striped -> raid0*
- clean up the takeover array
- allow lvconvert to actually call lv_raid_convert() on all takeover requests
in order to check parameters and display messages provided by takeover
functions rather than just "...not supported" from within lvconvert
- fix a typo
Resolves: rhbz1386148
If a device disappears after obtaining the list of devices but before
processing it as a member of that list, dmsetup exits with a failure code.
Most commands still produce what output they can in these circumstances,
but 'ls --tree' and 'info -c' with fields depending on device dependencies
didn't. Change this.
Keep for now function logic making its decision on string content.
We need bigger patch converting all things to bit-checks later.
This needs however bigger refactoring.
So this commit reverts some changes from:
c8b6c13015
Commit 088b3d036a allowed repair on cache origin RAID LVs
and restricted lvconvert actions on RAID SubLVs to change number of mirrors, repair,
replace and type changes in order to avoid unsuitable coversions on them.
This introduced a regression prohibiting --splitmirrors on any RAID SubLVs
(e.g. of cache or thin LVs; lvconvert-{cache,thin}-raid.sh tests failing).
Fix allows split mirrors again.
Fix some indenting whilst on it.
When we have already decoded arg_is_set into a local var
or already set segment type - already use these
values instead of repeating calls and string checks.
Seems some error path where not converted to 'new' ECMD return value.
Fix them to always 'goto out'.
Also drop unneeded 'ret = 0' when ret already is 0.
In case a RAID orig LV is being cached and fails, repair is impossible because
"lvconvert --repair" gets rejected.
Fix by allowing repair on cache orig RAID LVs and
"lvconvert --replace/--mirrors/--type {raid*|mirror|striped|linear}" as well.
Allow the same lvconvert actions on any cache pool and metadata RAID SubLVs.
Resolves: rhbz1380532
Make sure that the temporary dm_histogram used for the bounds
argument is freed in the case that the user provided a --bounds
argument on the command line.
The dm-raid target now rejects device rebuild requests during ongoing
resynchronization thus causing 'lvconvert --repair ...' to fail with
a kernel error message. This regresses with respect to failing automatic
repair via the dmeventd RAID plugin in case raid_fault_policy="allocate"
is configured in lvm.conf as well.
Previously allowing such repair request required cancelling the
resynchronization of any still accessible DataLVs, hence reasoning
potential data loss.
Patch allows the resynchronization of still accessible DataLVs to
finish up by rejecting any 'lvconvert --repair ...'.
It enhances the dmeventd RAID plugin to be able to automatically repair
by postponing the repair after synchronization ended.
More tests are added to lvconvert-rebuild-raid.sh to cover single
and multiple DataLV failure cases for the different RAID levels.
- resolves: rhbz1371717
Reload of thin-pool origin_only is designed to only post messages
to a thin-pool. It's not intended to be used for reload of thin-pool
table. Fix it by using standard call 'lv_update_and_reload()'.
Reinstantiate reporting of metadata percent usage for cache volumes.
Also show the same percentage with hidden cache-pool LV.
This regression was caused by optimization for a single-ioctl in
2.02.155.
Allow RAID scrubbing on cache origin sub-LV
This patch adds the ability to perform RAID scrubbing on the cache
origin sub-LV (https://bugzilla.redhat.com/1169495). Cache origin
operations are restricted to non-clustered RAID LVs until there can
be further testing in a cluster (even for exclusive activation).
User can either specify directly _corig LV
or he can specify cache LV and operation --syncation is
passed ONLY to _corig LV.
If users wants to manipulation with cache-pool devices - he
needs to specify this object name.
Signed-off-by: Jonathan Brassow <jbrassow@redhat.com>
Introduce 'hard limit' for max number of cache chunks.
When cache target operates with too many chunks (>10e6).
When user is aware of related possible troubles he
may increase the limit in lvm.conf.
Also verbosely inform user about possible solution.
Code works for both lvcreate and lvconvert.
Lvconvert fully supports change of chunk_size when caching LV
(and validates for compatible settings).
Commit e947c362dd introduced
config_settings.h file for central place to store all definitions for
config options. By mistake, it used report/colums_as_rows instead
of report/columns_as_rows (missing "n" in "columns").
If the number of stripes requested is incompatible with the requested
type of raid, give an error instead of adjusting it.
If no stripes argument is supplied, continue to use an appropriate
default.
a579ba2ac2 fixed a regression causing a segfault if no external
origin existed but broke the logic leading to erroneous error
messages and creations of split off exported VGs in case the
external origin and the pool LVs were allocated on different PVs.
- resolves rhbz1367459
Creating a RaidLV in VGs with very small extent sizes caused
late failure in the kernel giving a not very informative error
message. Catch the attempt early and display failure message
'Unable to create RAID LV: requires minimum VG extent size 4.00 KiB'.
- resoves rhbz1179970
'pvmove -n name pv1 pv2' called with the name of a top-level LV
failed with mentioned commit.
Enhance pvmove-raid-segtypes.sh to test for prohibited RAID SubLV moves.
'pvmove -n name pv1 pv2' allows to collocate multiple RAID SubLVs
on pv2 (e.g. results in collocated raidlv_rimage_0 and raidlv_rimage_1),
thus causing loss of resilence and/or performance of the RaidLV.
Fix this pvmove flaw leading to potential data loss in case of PV failure
by preventing any SubLVs from collocation on any PVs of the RaidLV.
Still allow to collocate any DataLVs of a RaidLV with their sibling MetaLVs
and vice-versa though (e.g. raidlv_rmeta_0 on pv1 may still be moved to pv2
already holding raidlv_rimage_0).
Because access to the top-level RaidLV name is needed,
promote local _top_level_lv_name() from raid_manip.c
to global top_level_lv_name().
- resolves rhbz1202497
Enforce mirror/raid0/1/10/4/5/6 type specific maximum images when
creating LVs or converting them from mirror <-> raid1.
Document those maxima in the lvcreate/lvconvert man pages.
- resolves rhbz1366060
Some settings are not suitable for override in interactive/shell
mode because such settings may confuse the code and it may end
up with unexpected behaviour. This is because of the fact that
once we're in the interactive/shell mode, we have already applied
some settings for the shell itself and we can't override them
further because we're already using those settings to drive the
interactive/shell mode. Such settings would get ignored silently
or, in worse case, they would mess up the existing configuration.
When lvm commands are executed in lvm shell, we cover the whole lvm
command execution within this shell now. That means, all messages logged
and status caught during each command execution is now recorded in the
log report, including overall command's return code.
We may call arg_count/grouped_arg_count/arg_value soon enough that
cmd->arg_values is not set yet.
Normally, when running a command, we execute lvm_run_command which in
turn calls _process_command_line to allocate and parse the command line
values and stores them in cmd->arg_values.
However, if we run lvm shell, this one doesn't accept any command line
options and we parse the command line for each command that is executed
within the lvm shell then. If we used any code that tries to access
cmd->arg_values through any of the the arg handling functions too
early, we could end up with a segfault due to uninitialized (NULL)
cmd->arg_values.
This patch just saves extra checks in all the code where arg handling
may be called too early so that the cmd->arg_values is not set up yet.
This does not apply to any of existing code, but subsequent patches
will need that.
With patches that will follow, this will make it possible to widen log
report coverage when commands are executed from lvm shell so the amount
of messages that may end up in stderr/stdout instead of log report are
minimized.
Add new log_context=shell and with log_object_type=cmd and
log_object_name=<command_name> for command log report to collect
overall return code from last command (this is reported under
log_type=status).
Currently, the output is separated in 3 parts and each part can go into
a separate and user-defined file descriptor:
- common output (stdout by default, customizable by LVM_OUT_FD environment variable)
- error output (stderr by default, customizable by LVM_ERR_FD environment variable)
- report output (stdout by default, customizable by LVM_REPORT_FD environment variable)
For example, each type of output goes to different output file:
[0] fedora/~ # export LVM_REPORT_FD=3
[0] fedora/~ # lvs fedora vg/abc 1>out 2>err 3>report
[0] fedora/~ # cat out
[0] fedora/~ # cat err
Volume group "vg" not found
Cannot process volume group vg
[0] fedora/~ # cat report
LV VG Attr LSize Layout Role CTime
root fedora -wi-ao---- 19.00g linear public Wed May 27 2015 08:09:21
swap fedora -wi-ao---- 500.00m linear public Wed May 27 2015 08:09:21
Another example in LVM shell where the report goes to "report" file:
[0] fedora/~ # export LVM_REPORT_FD=3
[0] fedora/~ # lvm 3>report
(in lvm shell)
lvm> vgs
(content of "report" file)
[1] fedora/~ # cat report
VG #PV #LV #SN Attr VSize VFree
fedora 1 2 0 wz--n- 19.49g 0
(in lvm shell)
lvm> lvs
(content of "report" file)
[1] fedora/~ # cat report
VG #PV #LV #SN Attr VSize VFree
fedora 1 2 0 wz--n- 19.49g 0
LV VG Attr LSize Layout Role CTime
root fedora -wi-ao---- 19.00g linear public Wed May 27 2015 08:09:21
swap fedora -wi-ao---- 500.00m linear public Wed May 27 2015 08:09:21
The MD raid6 personality being used to drive lvm raid6 LVs does
read-modify-write updates to any stripes and thus relies on correct
P and Q Syndromes being written during initial synchronization or
it may fail reconstructing proper user data in case of SubLVs failing.
We may not allow the '--nosync' option on
creation of raid6 LVs for that reason.
Update/fix 'man lvcreate' in that regard.
add lvcreate-raid-nosync.sh test script.
- Resolves rhbz1358532
We don't need to refresh whole cmd context if we drop profile after
processing LVM command - just like we don't refresh cmd context when
we're applying the profile. It's because profiles contain only safe
subset of settings which do not require complete cmd context refresh.
This patch calls process_profilable_config instead of
refresh_toolcontext if there was profile applied for the LVM
command only, not --config which requires toolcontext refresh.
The process_profilable_config just sets proper values based on
values of profilable settings, but it does not do complete
reinitialization of various parts (e.g. filters, logging etc.).
'lvchange --resync LV' or 'lvchange --syncaction repair LV' request the
RAID layout specific parity blocks in raid4/5/6 to be recreated or the
mirrored blocks to be copied again from the master leg/copy for raid1/10,
thus not allowing a rebuild of a particular PV.
Introduce repeatable option '--[raid]rebuild PV' to allow to request
rebuilds of specific PVs in a RaidLV which are known to contain corrupt
data (e.g. rebuild a raid1 master leg).
Add test lvchange-rebuild-raid.sh to test/shell doing rebuild
variations on raid1/10 and 5; add aux function check_status_chars
to support the new test.
- Resolves rhbz1064592
Prepare for new segment type conversion functionality in cases that
currently fail. In the short-term, we need to do this while limiting
the changes to the code paths for the conversions that are already
supported.
on any thin snap external origin LV which caused a segfault
when none existed as exposed by the vgsplit-thin.sh test.
Only call lv_is_on_pvs() if an external origin LV actually
exists and correct the related splitting logic.
General RAID and RAID segment type specific checks are added
to merge.c. New static _check_raid_seg() is called on each segment
of a RaidLV (which have just one) from check_lv_segments().
New checks caught some unititialized segment members
which are addressed here as well:
- initialize seg->region_size to 0 in lvcreate.c for raid0/raid0_meta
- initialize list seg->origin_list in lv_manip.c
Add matching support for -Z option also we doing full conversion
to cache-pool.
Extending coversion message to show which pool type is created
and whether the metadata will be wiped or remain unmodified.
Follow-up to 27a767d5e8.
Tunning behavior in a way we always prompt when option --zero is NOT specified.
Without -Z lvm expects user wants to 'reset' cache-pool metadata
(they could have been splitted from some cached LV)
If user doesn't want to zero metadata he needs to specify -Zn.
User may also avoid prompting for zeroing by using -Zy for
cache-pool (basically equals using --yes without -Z being given)
(unlike full convert case, there is no cache-pool being converted,
so there is not 'uncoditional' prompt in this case).
When volume was lvconvert-ed to a thin-volume with external origin,
then in case thin-pool was in non-zeroing mode
it's been printing WARNING about not zeroing thin volume - but
this is wanted and expected - so nothing to warn about.
So in this particular use case WARNING needs to be suppressed.
Adding parameter support for lvcreate_params.
So now lvconvert creates 'normal thin LV' in read-only mode
(so any read will 'return 0' for a moment)
then deactivate regular thin LV and reacreate in 'final R/RW' mode
thin LV with external origin and activate again.
When cache pool is reused for a new cached volume, there is
normally no need to 'keep' old cache-pool metadata as this
could cause major data lose.
Unlike with 'lvcreate -H -LX --cachepool' conversion, this lvconvert
path left the metadata unzeroed - partly for making easier some
debugging, but this was rather a bug.
So to keep possible reattach of 'unzeroed' metadata, user
now has to use 'lvconvert -Zn' for such conversion. In this case
the prompt will appear about possibe data loss and to proceed,
user has to confirm such operation. Without -Zn metadata are wiped.
Commit 3928c96a37 introduced
new defaults for raid number of stripes, which may cause
backwards compatibility issues with customer scripts.
Adding configurable option 'raid_stripe_all_devices' defaulting
to '0' (i.e. off = new behaviour) to select the old behaviour
of using all PVs in the VG or those provided on the command line.
In case any scripts rely on the old behaviour, just set
'raid_strip_all_devices = 1'.
- resolves rhbz1354650
The --uuid, --major and --alldevices arguments were incorrectly tested
after confirming argc is > 0, in a branch that only executes if argc
== 0 (i.e. they were unreachable).
Move all device checks before the test for argc and log an appropriate
error before returning.
raid0/raid0_meta type LVs don't have a default number of stripes when
created without '-i/--stripes Stripes' whereas other raid types have one.
Patch sets the default for raid0/raid0_meta to 2 stripes.
The default amount of stripes for raid4/5/10 is changed to 2 and for raid6 to 3
rather than using all PVs in the VG or those provided on the command line.
This is to avoid unintended high number of stripes in case of many PVs.
To select a different amount of stripes from the default,
use 'lvcreate -i/--stripes Stripes'.
- resolves rhbz1354650
Resync attempts on raid0/raid0_meta via 'lvchange --resync ...'
cause segfaults.
'lvchange --syncaction ...' doesn't get rejected either.
Prohibit both on raid0/raid0_meta LVs.
- resolves rhbz1354656
4420d41fea introduced recursive split of lvs which
splits a top-level LV together with it's sub LVs.
This lead to invalid temporary list pointers
causing hangs/OOM situations.
Patch updates the temporary list pointer
referencing a moved sub LV.
- resolves rhbz1354686
Make the --filemap switch take no arguments and instead accept one
or more files on the command line to be mapped and placed into
groups.
This allows --filemap to be used with a glob:
# dmstats create --filemap *
rhel5.10-1.qcow2: Created new group with 87 region(s) as group ID 1564.
rhel5.10.qcow2: Created new group with 8 region(s) as group ID 1651.
rhel7.0-1.qcow2: Created new group with 11 region(s) as group ID 1659.
rhel7.0.qcow2: Created new group with 1454 region(s) as group ID 1670.
vm.img: Created new group with 2 region(s) as group ID 3124.
lvconvert --splitcache VG/CachePool_corig
Allow the split via the hidden/used cache pool for the time being,
since the new lvconvert code did intend to allow it, but was just
missing the exception in the list of hidden LVs that were allowed.
The preferred method for splitcache is to run it on the visible
cache LV, not the hidden cache pool. That may eventually become
the only method since we try to avoid running commands on
hidden LVs.
When a 'dmstats create --filemap' operation fails (e.g. during
open(2), close(2), or dm_stats_create_regions_from_fd()), use the
canonical version of the path. This avoids cryptic/confusing error
messages when symbolic links exist in the path argument given:
# findmnt /var/lib/libvirt/images -otarget,source
TARGET SOURCE
/var/lib/libvirt/images /dev/mapper/vg_hex-lv_images
# readlink /var/lib/libvirt/images/my.img
/boot/my.img
# dmstats create --filemap /var/lib/libvirt/images/my.img
Cannot map file: not a device-mapper device.
Could not create regions from file /var/lib/libvirt/images/my.img
Command failed
Using the canonical path the error is immediately obvious:
# dmstats create --filemap /var/lib/libvirt/images/my.img
Cannot map file: not a device-mapper device.
Could not create regions from file /boot/my.img
Command failed
Grouping is also useful in combination with --segments: creating a
group allows both individual segment data and data for the device
as a whole to be presented in the same report.
Support grouping for 'create --segments' in the same manner as for
'create --filemap'; group regions by default, applying an optional
alias specified with --alias, unless the user specifies --nogroup.
Add a new option to the create command to create regions that map the
extents of a file:
# dmstats create --filemap /path/to/file
/path/to/file: Created new group with 10 region(s) as group ID 0.
When performing a --filemap no device argument is required (and
supplying one results in error) since the device to bind to is implied
by the file path and is obtained directly from an fstat().
Grouping may be optionally disabled by the --nogroup switch: in this
case the command will report each region individually:
# dmstats create --nogroup --filemap /path/to/file
/path/to/file: Created new region with 1 area as region ID 0.
/path/to/file: Created new region with 1 area as region ID 1.
/path/to/file: Created new region with 1 area as region ID 2.
When grouping regions the group alias is automatically set to the
basename (as returned by dm_basename()) of the provided file.
This can be overridden to a user-defined value at the command line by
use of the --alias option.
If grouping is disabled no alias can be set.
Use of offset and subdivision options (--start, --length, --segments,
--areas, --areasize).
Setting aux_data and histograms for groups is possible but is not
currently implemented.
The code could perform this conversion but ironically
did not recognize the standard command form, only the
the unpreferred "implication-based" command form.
"lvconvert --type linear VG/RaidLV" would fail, but
"lvconvert --mirrors 0 VG/RaidLV" would succeed.
The code could perform this conversion but ironically
did not recognize the standard command form, only the
the unpreferred "implication-based" command form.
"lvconvert --type linear VG/MirrorLV" would fail, but
"lvconvert --mirrors 0 VG/MirrorLV" would succeed.
Make it clear that the "aux data" presented in reports is the user
data stored in the field (and does not include any library-internal
state such as group descriptors) by renaming the field to user_data
and changing the heading to "UserData".
Replace --statstype=area,region,group with a separate switch for
each object type: --area, --region, --group. Omitting any object
type switch will use the defaults for the current command (regions
and groups for list, and regions, groups and areas for verbose list).
Replace the 'name' field with 'statsname' in order to report alias
names for groups, and include the 'group_id' field between statsname
and the 'region_id' field to make it clear to the user when groups
are in use.
Walk avaiable groups and regions (in addition to areas) and report
aggregate statistics and properties.
A new switch is added to filter the type of obects inclued in the
report:
--statstype={all,area,region,group}
The type of the current row is also available in a new
DR_STATS_META field 'type'.
To allow the names used to describe statistics report objects to
change (for e.g to support groups and region and group aliases)
introduce a new "stats_name" field that evaluates to the correct
name for the object being reported.
Add a pair of commands to create and delete stats groups:
dmstats group --regions REGIONS
dmstats ungroup --groupid ID
REGIONS specifies a list of regions to be included in the group.
Regions are specified as a comma separated list in order of
increasing region ID. Ranges may be specified as a hypen separated
pair of values giving the first and last member of the range.
With a single report (--count=1) no timerfd is set up and the cycle
and current timestamps should be freed during the single call to
_update_interval_times().
Add new logic to identify each unique operation and route
it to the correct function to perform it. The functions
that perform the conversions remain unchanged.
This new code checks every allowed combination of LV type
and requested operation, and for each valid combination
calls the function that performs that conversion.
The first stage of option validation which checks for
incompatible combinations of command line options, is done
done before process_each is called. This is unchanged.
(This new code will allow that first stage validation to
be simplified in a future commit.)
The second stage of checking options against the specific
LV type is done by this new code. For each valid combination
of operation + LV type, the new code calls an existing
function that implements it.
With this in place, the ad hoc checks for valid combinations
of LV types and operations can be removed from the existing
code in a future commit.
(The #if 0 is used to keep the patch clean, and the
disabled code will be removed by a following patch.)
Previously, a command sent lvmetad new VG metadata in vg_commit().
In vg_commit(), devices are suspended, so any memory allocation
done by the command while sending to lvmetad, or by lvmetad while
updating its cache could deadlock if memory reclaim was triggered.
Now lvmetad is updated in unlock_vg(), after devices are resumed.
The new method for updating VG metadata in lvmetad is in two phases:
1. In vg_write(), before devices are suspended, the command sends
lvmetad a short message ("set_vg_info") telling it what the new
VG seqno will be. lvmetad sees that the seqno is newer than
the seqno of its cached VG, so it sets the INVALID flag for the
cached VG. If sending the message to lvmetad fails, the command
fails before the metadata is committed and the change is not made.
If sending the message succeeds, vg_commit() is called.
2. In unlock_vg(), after devices are resumed, the command sends
lvmetad the standard vg_update message with the new metadata.
lvmetad sees that the seqno in the new metadata matches the
seqno it saved from set_vg_info, and knows it has the latest
copy, so it clears the INVALID flag for the cached VG.
If a command fails between 1 and 2 (after committing the VG on disk,
but before sending lvmetad the new metadata), the cached VG retains
the INVALID flag in lvmetad. A subsequent command will read the
cached VG from lvmetad, see the INVALID flag, ignore the cached
copy, read the VG from disk instead, update the lvmetad copy
with the latest copy from disk, (this clears the INVALID flag
in lvmetad), and use the correct VG metadata for the command.
(This INVALID mechanism already existed for use by lvmlockd.)
Simplify code around _do_get_report_selection - remove "expected_idxs[]"
argument which is superfluous and add "allow_single" switch instead to
allow for recognition of "--configreport <report_name> -S" as well as
single "-S" if needed.
Null pointer dereferences (FORWARD_NULL) /safe/guest2/covscan/LVM2.2.02.158/tools/reporter.c: 961 in _do_report_get_selection()
Null pointer dereferences (FORWARD_NULL) Dereferencing null pointer "single_args".
Uninitialized variables (UNINIT) /safe/guest2/covscan/LVM2.2.02.158/tools/toollib.c: 3520 in _process_pvs_in_vgs()
Uninitialized variables (UNINIT) Using uninitialized value "do_report_ret_code".
Null pointer dereferences (REVERSE_INULL) /safe/guest2/covscan/LVM2.2.02.158/libdm/libdm-report.c: 4745 in dm_report_output()
Null pointer dereferences (REVERSE_INULL) Null-checking "rh" suggests that it may be null, but it has already been dereferenced on all paths leading to the check.
Incorrect expression (MISSING_COMMA) /safe/guest2/covscan/LVM2.2.02.158/lib/log/log.c: 280 in _get_log_level_name()
Incorrect expression (MISSING_COMMA) In the initialization of "log_level_names", a suspicious concatenated string ""noticeinfo"" is produced.
Null pointer dereferences (FORWARD_NULL) /safe/guest2/covscan/LVM2.2.02.158/tools/reporter.c: 816 in_get_report_options()
Null pointer dereferences (FORWARD_NULL) Comparing "mem" to null implies that "mem" might be null.
Use common API design and pass just LV pointer to lv_manip.c functions.
Read cmd struct via lv->vg->cmd when needed.
Also do not try to return EINVALID_CMD_LINE error when we
have already openned VG - this error code can only be returned before
locking VG.
In commit 6ae22125, vgcfgrestore began disabling lvmetad
while running, and rescanned to enable it again at the end,
but missed the rescanning/enabling in the error case.
Previously, vgcfgrestore would attempt to vg_remove the
existing VG from lvmetad and then vg_update to add the
restored VG. But, if there was a failure in the command
or with vg_update, the lvmetad cache would be left incorrect.
Now, disable lvmetad before the restore begins, and then
rescan to populate lvmetad from disk after restore has
written the new VG to disk.
Reporting commands can be of different types (even if the command name
is the same):
- pvs command can be either of PVS, PVSEGS or LABEL report type,
- vgs command is of VGS report type,
- lvs command is of LVS or SEGS report type.
Use basic report type when looking for report prefix used for
--configreport option.
This means that:
- 'pvs --configreport pv' applies to PVS, PVSEGS or LABEL report type
- 'vgs --configreport vg' applies to VGS report type
- 'lvs --configreport lv' applies to LVS and SEGS report type
This patch adds structures and functions to reroute error and warning
logs to log report, if it's set.
There are 5 new functions:
- log_set_report
Set log report where logging will be rerouted.
- log_set_report_context
Set context globally so any report_cmdlog call will use it.
- log_set_report_object_type
Set object type globally so any report_cmdlog call will use it.
- log_set_report_object_name_and_id
Set object ID and name globally so any report_cmdlog call will use it.
- log_set_report_object_group_and_group_id
Set object group ID and name globally so any report_cmdlog call will use it.
These functions will be called during LVM command processing so any logs
which are rerouted to log report contain proper information about current
processing state.
The lvm fullreport works per VG and as such, the vg, lv, pv, seg and
pvseg subreport is done for each VG. However, if the PV is not part of
any VG yet, we still want to display pv and pvseg subreports for these
"orphan" PVs - so enable this for lvm fullreport's process_each_vg call.
If we have fullreport, make sure that the options/sort keys used for
each report doesn't change its type - we want to preserve the original
type so it's always 5 different subreports within fullreport (vg, lv, pv,
seg, pvseg). Since we have all report types within fullreport, users
should add fields under proper subreport type - this minimizes
duplication of info displayed on output.
Groupable args (the ones marked with ARG_GROUPABLE flag) start a new
group of args if:
- this is the first time we hit such a groupable arg,
- or if non-countable arg is repeated.
However, there may be cases where we want to give priorities when
forming groups and hence force new group creation if we hit an arg
with higher grouping priority.
For example, let's assume (for now) hypothetical sequence of args used:
lvs -o lv_name --configreport log -o log_type --configreport lv -o +vg_name
Without giving any priorites, we end up with:
lvs -o lv_name --configreport log -o log_type --configreport lv -o +vg_name
| | | | | |
\__________GROUP1___________/ \________GROUP2___________/ \_GROUP3_/
This is because we hit "-o" as the first groupable arg. The --configreport,
even though it's groupable too, it falls into the previous "-o" group.
While we may need to give priority to the --configreport arg that should
always start a new group in this scenario instead:
lvs -o lv_name --configreport log -o log_type --configreport lv -o +vg_name
| | | | | |
\_GROUP1_/ \_________GROUP2___________/ \_________GROUP3__________/
So here "-o" started a new group but since "--configreport" has higher
priority than "-o", it starts fresh new group now and hence the rest of
the command line's args are grouped by --configreport now.
lvm fullreport executes 5 subreports (vg, pv, lv, pvseg, seg) per each VG
(and so taking one VG lock each time) within one command which makes it
easier to produce full report about LVM entities.
Since all 5 subreports for a VG are done under a VG lock, the output is
more consistent mainly in cases where LVM entities may be changed in
parallel.
Add any report (pvs/vgs/lvs) currently processed to current report group
which is part of processing handle and which already contains log
report. This way both log report and pvs/vgs/lvs report will be
reported as a whole within a group, thus having same output format as
selected by --reportformat option.
If there's parent processing handle, we don't need to create completely
new report group and status report - we'll just reuse the one already
initialized for the parent.
Currently, the situation where this matter is when doing internal report
to do the selection for processing commands where we have parent processing
handle for the command itself and processing handle for the selection
part (that is selection for non-reporting tools).
Wire up report group creation with log report in struct
processing_handle and call report_format_init during processing handle
initialization (init_processing_handle fn) and destroy it while
destroing processing handle (destroy_processing_handle fn).
This way, all the LVM command processing using processing handle
has access to log report via which the current command log
can be reported as items are processed.
Separating common report and per-report arguments prepares the code for
handling several reports per one command (for example, the command log
report and LVM command report itself).
Each report can have sort keys, options (fields), list of fields to
compact and selection criteria set individually. Hooks for setting these
per report within one command will be a part of subsequent patches, this
patch only separates new struct single_report_args out of existing
struct report_args.
New report/output_format configuration sets the output format used
for all LVM commands globally. Currently, there are 2 formats
recognized:
- basic (the classical basic output with columns and rows, used by default)
- json (output is in json format)
Add new --reportformat option and new report_format_init function that
checks this option and creates new report group accordingly, also
preparing log report handle and adding it to the report group just
created.
This is a preparation for new CMDLOG report type which is going to be
used for reporting LVM command log.
The new report type introduces several new fields (log_seq_num, log_type,
log_context, log_object_type, log_object_group, log_object_id, object_name,
log_message, log_errno, log_ret_code) as well as new configuration settings
to set this report type (report/command_log_sort and report/command_log_cols
lvm.conf settings).
This patch also introduces internal report_cmdlog helper function
which is a wrapper over dm_report_object to report command log via
CMDLOG report type and which is going to be used throughout the code
to report the log items.
commit 15da467b was meant to address the case where
use_lvmetad=1 in lvm.conf, and lvmetad is not available,
in which case, pvscan --cache -aay should activate LVs.
But the commit unintentionally also changed the case
where use_lvmetad=0 in lvm.conf, in which case
pvscan --cache -aay should not activate LVs, so fix
that here.
When pvscan --cache -aay fails to connect to lvmetad it will
simply exit and do nothing. Change this so that it will
skip the lvmetad cache step and do the activation step from
disk.
A number of places are working on a specific dev when they
call lvmcache_info_from_pvid() to look up an info struct
based on a pvid. In those cases, pass the dev being used
to lvmcache_info_from_pvid(). When a dev is specified,
lvmcache_info_from_pvid() will verify that the cached
info it's using matches the dev being processed before
returning the info. Calling code will not mistakenly
get info for the wrong dev when duplicate devs exist.
This confusion was happening when scanning labels when
duplicate devs existed. label_read for the first dev
would add an info struct to lvmcache for that dev/pvid.
label_read for the second dev would see the pvid in
lvmcache from first dev, and mistakenly conclude that
the label_read from the second dev can be skipped
because it's already been done. By verifying that the
dev for the cached pvid matches the dev being read,
this mismatch is avoided and the label is actually read
from the second duplicate.
pvmove began processing tags unintentionally from commit,
6d7dc87cb pvmove: use toollib
pvmove works on a single PV, but tags can match multiple PVs.
If we allowed tags, but processed only the first matching PV,
then the resulting PV would be unpredictable.
Also, the current processing code does not allow us to simply
report an error and do nothing if more than one PV matches the tag,
because the command starts processing PVs as they are found,
so it's too late to do nothing if a second PV matches.
If configuration consists of several sources in config cascade
("config cascade" defined in man lvmconfig(8)), lvmconfig displayed
only difference from defaults of the topmost config in the cascade.
Fix lvmconfig to display complete difference, considering all
the configuration in the cascade.
For example, before this patch:
(use_lvmetad=0 set in lvm.conf which differs from defaults)
$ lvmconfig --type diff
global {
use_lvmetad=0
}
(compact_output=1 set on cmd line)
$ lvmconfig --type diff --config report/compact_output=1
report {
compact_output=1
}
(headings=0 set in profile)
$ lvmconfig --type diff --commandprofile test
report {
headings=0
}
(difference in topmost configuration source is displayed)
$ lvmconfig --type diff --commandprofile test --config report/compact_output=1
report {
compact_output=1
}
With this patch applied (the config cascade is merged before looking for
difference from defaults in configuration):
$ lvmconfig --type diff
global {
use_lvmetad=0
}
$ lvmconfig --type diff --config report/compact_output=1
report {
compact_output=1
}
global {
use_lvmetad=0
}
$ lvmconfig --type diff --profile test
report {
headings=0
}
global {
use_lvmetad=0
}
$ lvmconfig --type diff --profile test --config report/compact_output=1
report {
headings=0
compact_output=1
}
global {
use_lvmetad=0
}
In the same way that process_each_vg() can be passed
a single VG name to process, also allow process_each_lv()
to be passed a single VG name and LV name to process.
This refactors the code for autoactivation. Previously,
as each PV was found, it would be sent to lvmetad, and
the VG would be autoactivated using a non-standard VG
processing function (the "activation_handler") called via
a function pointer from within the lvmetad notification path.
Now, any scanning that the command needs to do (scanning
only the named device args, or scanning all devices when
there are no args), is done first, before any activation
is attempted. During the scans, the VG names are saved.
After scanning is complete, process_each_vg is used to do
autoactivation of the saved VG names. This makes pvscan
activation much more similar to activation done with
vgchange or lvchange.
The separate autoactivate phase also means that if lvmetad
is disabled (either before or during the scan), the command
can continue with the activation step by simply not using
lvmetad and reverting to disk scanning to do the
activation.
Add support for active cache LV.
Handle --cachemode args validation during command line processing.
Rework some lvm2 internal to use lvm2 defined CACHE_MODE enums
indepently on libdm defines and use enum around the code instead
of passing and comparing strings.
The _report fn is getting big - separate it in two:
- _report fn to get all the options and arguments
- _do_report fn for reporting itself
Also, place all the variables/arguments in one structure for easier
handling of the variables around.
Previously, duplicate PVs were processed as a side effect
of processing the "chosen" PV in lvmcache. The duplicate
PV would be hacked into lvmcache temporarily in place of
the chosen PV.
In the old way, we had to always process the "chosen" PV
device, even if a duplicate of it was named on the command
line. This meant we were processing a different device than
was asked for. This could be worked around by naming
multiple duplicate devs on the command line in which case
they were swapped in and out of lvmcache for processing.
Now, the duplicate devs are processed directly in their
own processing loop. This means we can remove the old
hacks related to processing dups as a side effect of
processing the chosen device. We can now simply process
the device that was named on the command line.
When the same PVID exists on two or more devices, one device
is preferred and used in the VG, and the others are duplicates
and are not used in the VG. The preferred device exists in
lvmcache as usual. The duplicates exist in a specical list
of unused duplicate devices.
The duplicate devs have the "d" attribute and the "duplicate"
reporting field displays "duplicate" for them.
'pvs' warns about duplicates, but the formal output only
includes the single preferred PV.
'pvs -a' has the same warnings, and the duplicate devs are
included in the output.
'pvs <path>' has the same warnings, and displays the named
device, whether it is preferred or a duplicate.
Wait to compare and choose alternate duplicate devices until
after all devices are scanned. During scanning, the first
duplicate dev is kept in lvmcache, and others are kept in a
new list (_found_duplicate_devs).
After all devices are scanned, compare all the duplicates
available for a given PVID and decide which is best.
If the dev used in lvmcache is changed, drop the old dev
from lvmcache entirely and rescan the replacement dev.
Previously the VG metadata from the old dev was kept in
lvmcache and only the dev was replaced.
A new config setting devices/allow_changes_with_duplicate_pvs
can be set to 0 which disallows modifying a VG or activating
LVs in it when the VG contains PVs with duplicate devices.
Set to 1 is the old behavior which allowed the VG to be
changed.
The logic for which of two devs is preferred has changed.
The primary goal is to choose a device that is currently
in use if the other isn't, e.g. by an active LV.
. prefer dev with fs mounted if the other doesn't, else
. prefer dev that is dm if the other isn't, else
. prefer dev in subsystem if the other isn't
If neither device is preferred by these rules, then don't
change devices in lvmcache, leaving the one that was found
first.
The previous logic for preferring a device was:
. prefer dev in subsystem if the other isn't, else
. prefer dev without holders if the other has holders, else
. prefer dev that is dm if the other isn't
Rather than doing repeated translations from name to
device when comparing args to existing PVs, do one
translation of the arg names and saving the device,
before checking existing PVs.
pvscan autoactivation has its own VG processing implementation,
so it can't properly handle things like foreign or shared VGs,
so make it ignore those VG types (or errors from them) as best
as possible.
Add a FIXME stating that pvscan autoactivation must really be
moved to the standard VG processing by calling process_each_vg
to do activation once the scanning / cache update is finished.
The lvmetad connection is created within the
init_connections() path during command startup,
rather than via the old lvmetad_active() check.
The old lvmetad_active() checks are replaced
with lvmetad_used() which is a simple check that
tests if the command is using/connected to lvmetad.
The old lvmetad_set_active(cmd, 0) calls, which
stopped the command from using lvmetad (to revert to
disk scanning), are replaced with lvmetad_make_unused(cmd).
Commands already check if the lvmetad token is valid,
and if not, they rescan devices to repopulate lvmetad
before running. Now, in addition to checking the
lvmetad token, they also check if the lvmetad disabled
flag is set. If so, they do not use the lvmetad cache
and revert to disk scanning.
process_each_pv was doing:
1. lvmcache_seed_infos_from_lvmetad()
sends pv_list request to lvmetad.
2. get_vgnameids()
sends vg_list request to lvmetad.
3. _get_all_devices()
first calls lvmcache_seed_infos_from_lvmetad(),
which is a no-op if it's already been called.
Because get_vgnameids() does not use the information
from lvmcache_seed_infos_from_lvmetad(), it does not
need to be called prior to get_all_devices where
it is actually needed.
Commit 971ab733b7 ("thin: activation of
merging thin snapshot") also added an incorrect deactivation attempt
for non-thin LVs: find_snapshot(lv)->lv is not designed to be
activated and any attempt to deactivate it is incorrect.
Move checking the lvmetad state, and the possible rescan,
out of lvmetad_send() to the start of the command.
Previously, the token mismatch and rescan would occur
within lvmetad_send() for some other request. Now,
the token mismatch is detected earlier, so the
rescan can be done before the main command is in
progress. Rescanning deep within the processing of
another command will disturb the lvmcache state of
that other command.
A rescan already exists at the start of the command
for the case where foreign VGs are going to be read.
This same rescan is now also performed when there is
an lvmetad token mismatch (from a changed global_filter).
The commands pvscan/vgscan/lvscan/vgimport are excluded
from this preemptive checking/rescanning for lvmetad
because they want to do rescanning themselves explicitly.
If rescanning devices fails, then lvmetad has not been
correctly repopulated and should not be used, so make
the command revert to not using lvmetad.
We don't have any report field of this type yet. Return this patch into
the play if we really need that. Currenly we always report status
(result of "status" dm ioctl) for an LV as a whole where we choose
segment which represents the LV, not calling status for each possible
segment it contains - we don't need this now so I'm removing it to
not make the code more complex uselessly.
This is hotfix for RHBZ: https://bugzilla.redhat.com/1324537
However already the %FREE is not a good fit and we need something
better. Meanwhile make -l%PVS work at least as good as %FREE
for thin-pool.
TODO: this needs rework - it should be allocator to do all the size
decisions at one place.
Allow pvchange and pvresize to process exported VGs,
and have them check for the exported state in their
single function.
Previously, the exported VG state would trigger a
failure in vg_read()/ignore_vg() because the VGs are
being read with READ_FOR_UPDATE. Because these commands
read all VGs to search for the intended PVs, any
exported VG would trigger a failure, even if it was
not related to the intended PV.
Use <> around user entered option parameters to make it visually
different (just like we use Italic style in man page).
TODO:
In the future we should consistently provide such notation and
possibly generate it automagically from some internal data structures.
Preferably for man pages as well so we report actual set of supported
options.
Commit abd9618dd8 tried to improve
parsing of vg name from logical path - but still missed couple
corner cases.
This patch further improves the logic and reuses
validate_lvname_param() for parsing of lv_name.
Also explicitly checks for LVM_VG_NAME in the right case.
So now also properly parses cases like:
'lvconvert --repairt vg/'
and will provide correct error message.
With the recent conversion of pvcreate/pvremove to the
common toollib processing function, skipping in-use PVs
in _process_pvs_in_vg prevented them from being protected
as intended by the in-use flag.
The processing code for pvcreate/pvremove checks for the
in-use state itself and prevents using an in-use PV.
If a PV is skipped, it looks like an unused device and
is not protected from being used in pvcreate/pvremove.
This command option can be used to trigger a D-Bus
notification independent of the usual notifications
that are sent from other commands as an effect of
changes to PV/VG/LV state. If lvm is not built with
dbus notification support or if notify_dbus is disabled
in the config, this command will exit with an error.
When a command modifies a PV or VG, or changes the
activation state of an LV, it will send a dbus
notification when the command is finished. This
can be enabled/disabled with a config setting.
The metadata/record_lvs_history is global switch which enables or
disables recording historical LVs in metadata.
If both metadata/record_lvs_history=1 lvm.conf option and
--nohistory command switch is used at the same time, the
--nohistory prevails.
When processing LVs in a VG and when the -H|--history switch is used,
make process_each_lv_in_vg to iterate over historical volumes too.
For each historical LV, we use dummy struct logical_volume instance with
the "this_glv" reference set to a wrapper over proper struct
historical_logical_volume representation. This makes it possible to process
historical LVs just like normal live LVs (though a dummy one without any
segments and all the other fields zeroed and blank) and it also allows
for using all historical LV related information via lv->this_glv->historical
reference.
One can use a simple call to lv_is_historical to make a difference between
live and historical LV in all the code that is called by process_each_* fns.
This patch adds "include_historical_lvs" field to struct cmd_context to
make it possible for the command to switch between original funcionality
where no historical LVs are processed and functionality where historical
LVs are taken into account (and reported or processed further). The switch
between these modes is done using the '-H|--history' switch on command
line.
The include_historical_lvs state is then passed to process_each_* fns
using the "include_historical_lvs" field within struct processing_handle.
Add new structures and new fields in existing structures to support
tracking history of LVs (the LVs which don't exist - the have been
removed already):
- new "struct historical_logical_volume"
This structure keeps information specific to historical LVs
(historical LV is very reduced form of struct logical_volume +
it contains a few specific fields to track historical LV
properties like removal time and connections among other LVs).
- new "struct generic_logical_volume"
Wrapper for "struct historical_logical_volume" and
"struct logical_volume" to make it possible to handle volumes
in uniform way, no matter if it's live or historical one.
- new "struct glv_list"
Wrapper for "struct generic_logical_volume" so it can be
added to a list.
- new "indirect_glvs" field in "struct logical_volume"
List that stores references to all indirect users of this LV - this
interconnects live LV with historical descendant LVs or even live
descendant LVs.
- new "indirect_origin" field in "struct lv_segment"
Reference to indirect origin of this segment - this interconnects
live LV (segment) with historical ancestor.
- new "this_glv" field in "struct logical_volume"
This references an existing generic_logical_volume wrapper for this
LV, if used. It can be NULL if not needed - which means we're not
handling historical LVs at all.
- new "historical_lvs" field in "struct volume group
List of all historical LVs read from VG metadata.
Since we want to read env LVM_VG_NAME vg names,
we cannot just check LV names which do contain '/'.
So before the patch commands like:
> lvconvert --repair vg
Before:
Please provide a valid volume group name
After:
Path required for Logical Volume "vg".
Please provide a valid volume group name
> LVM_VG_NAME=vg lvconvert --repair vg
Before:
Please provide a valid volume group name
After:
Can't find LV vg in VG vg
Commit ca878a3426 introduced an issue
that zero sized extesion suddenly started to be accepted and
missed to return error.
Properly check invalid input values for sizes.
"pvcreate_each_params" was a temporary name used
to transition from the old "pvcreate_params".
Remove the old pvcreate_params struct and rename the
new pvcreate_each_params struct to pvcreate_params.
Rename various pvcreate_each_params terms to simply
pvcreate_params.
Use the new pvcreate_each_device() function from
toollib, previously added for pvcreate, in place
of the old pvcreate_vol().
This also requires shifting the location where the
lock is acquired for the new VG name. The lock for
the new VG is supposed to be acquired before pvcreate.
This means splitting the vg_lock_newname() out of
vg_create(), and calling vg_lock_newname() directly
before pvcreate, and then calling the remainder of
vg_create() after pvcreate.
The new function vg_lock_and_create() now does
vg_lock_newname() + vg_create(), like the previous
version of vg_create().
The lock on the new VG name is released before the
pvcreate and reacquired after the pvcreate because
pvcreate needs to reset lvmcache, which doesn't work
when locks are held. An exception could likely be
made for the new VG name lock, which would allow
vgcreate to hold the new VG name lock across the
pvcreate step.
This is common code for handling PV create/remove
that can be shared by pvcreate/vgcreate/vgextend/pvremove.
This does not change any commands to use the new code.
- Pull out the hidden equivalent of process_each_pv
into an actual top level process_each_pv.
- Pull the prompts to the top level, and do not
run any prompts while locks are held.
The orphan lock is reacquired after any prompts are
done, and the devices being created are checked for
any change made while the lock was not held.
Previously, pvcreate_vol() was the shared function for
creating a PV for pvcreate, vgcreate, vgextend.
Now, it will be toollib function pvcreate_each_device().
pvcreate_vol() was called effectively as a helper, from
within vgcreate and vgextend code paths.
pvcreate_each_device() will be called at the same level
as other process_each functions.
One of the main problems with pvcreate_vol() is that
it included a hidden equivalent of process_each_pv for
each device being created:
pvcreate_vol() -> _pvcreate_check() ->
find_pv_by_name() -> get_pvs() ->
get_pvs_internal() -> _get_pvs() -> get_vgids() ->
/* equivalent to process_each_pv */
dm_list_iterate_items(vgids)
vg = vg_read_internal()
dm_list_iterate_items(&vg->pvs)
pvcreate_each_device() reorganizes the code so that
each-VG-each-PV loop is done once, and uses the standard
process_each_pv function at the top level of the function.
This uses the vg->pv_write_list in place of the
vg->pvs_to_write list, and eliminates the use of
pvcreate_params. The label remove and zeroing
steps are shifted out of vg_write() to the higher
level like pvcreate will do.
If we know that a PV belongs to some VG and we're missing metadata
(because we have only those PV(s) from VG present in the system that
don't have metadata areas), we should skip such PV when processing
under system ID.
This is because we know that the PV belongs to some VG, but we
really can't decide whether it matches system ID unless the VG
metadata is present again.
The backup_restore_vg is used directly for restoring the VG from backup.
It's also used to do the VG conversions from one metadata format to
another which means vgconvert calls backup_restore_vg too.
When restoring VG from backup, we need to rewrite/write PV headers as
PVs may have been orphans before and now they're becoming part of some
VG - we need to write the PV_EXT_USED flag at least.
When using the backup_restore_vg for vgconvert, we need to write
completely new PV header in different format.
Avoid the special "pv_write" call and handling that was used before
this patch in vgconvert (vgconvert_single function to be more precise)
and reuse existing internal interface to register PV header for writing
(or rewriting) via vg->pvs_to_write list instead like we do it elsewhere
in the code.
This patch also resolves a problem in which PV headers with target
format were written in the vgconvert_single fn as orphans and VG
metadata were added later on - this was a tiny hack actually.
We can't do this now - we need to write the PV as belonging
to a VG because otherwise the PV_EXT_USED flag won't be written
properly (if the PV header is written as orphan, the PV_EXT_USED
is set to 0, of course, even though metadata are attached later).
So this patch removes this tiny inconsistency which was passing
just fine before because we didn't have any relation to the VG
in PV header before. Now we have the PV_EXT_USED flag which says
the "PV is used in some VG".
If we know that the PV is orphan, meaning there's at least one MDA on
that PV which does not reference any VG and at the same time there's
PV_EXT_USED flag set, we're certainly in an inconsistent state and we
need to fix this.
For example, such situation can happen during vgremove/vgreduce if we
removed/reduced the VG, but we haven't written PV headers yet because
vgremove stopped abruptly for whatever reason just before writing new
PV headers with updated state, including PV extension flags (and so the
PV_EXT_USED flag).
However, in case the PV has no MDAs at all, we can't double-check
whether the PV_EXT_USED is correct or not - if that PV is marked
as used, it's either:
- really used (but other disks with MDAs are missing)
- or the error state as described above is hit
User needs to overwrite the PV header directly if it's really clear
the PV having no MDAs does not belong to any VG and at the same time
it's still marked as being in use (pvcreate -ff <dev_name> will fix this).
For example - /dev/sda here has 1 MDA, orphan and is incorrectly marked
with PV_EXT_USED flag:
$ pvs --binary -o+pv_in_use
WARNING: Found inconsistent standalone Physical Volumes.
WARNING: Repairing flag incorrectly marking Physical Volume /dev/sda as used.
PV VG Fmt Attr PSize PFree InUse
/dev/sda lvm2 --- 128.00m 128.00m 0
Make sure we won't use a PV that is already marked as used. Normally,
VG metadata would stop us from doing that, but we can run into a
situation where such metadata is missing because PVs with MDAs
are missing and the PVs left are the ones with 0 MDAs.
(/dev/sda in this example has 0 MDAs and it belongs to a VG,
but other PVs with MDA are missing)
$ pvs -o pv_name,pv_mda_count /dev/sda
PV #PMda
/dev/sda 0
$ pvcreate /dev/sda
PV '/dev/sda' is marked as belonging to a VG but its metadata is missing.
Can't initialize PV '/dev/sda' without -ff.
$ pvchange -u /dev/sda
PV '/dev/sda' is marked as belonging to a VG but its metadata is missing.
Can't change PV '/dev/sda' without -ff.
Physical volume /dev/sda not changed
0 physical volumes changed / 1 physical volume not changed
$ pvremove /dev/sda
PV '/dev/sda' is marked as belonging to a VG but its metadata is missing.
(If you are certain you need pvremove, then confirm by using --force twice.)
$ vgcreate vg /dev/sda
Physical volume '/dev/sda' is marked as belonging to a VG but its metadata is missing.
Unable to add physical volume '/dev/sda' to volume group 'vg'.
This is a hotfix for a bug introduced in
6d7dc87cb3.
The bug description: First we allocate memory for
processing handle (at an address 1) then we
allocate some memory on the same pool for later use
in pvmove_poll function inside the process_each_pv
function (at an address 2). After we jump out of
process_each_pv we called destroy_processing_handle.
As a result of destroying the handle memory pool could
deallocate all memory at address 1 or higher. The
pvmove_poll function tried to copy a memory allocated
at address 2 that could be returned to the system.
If it was so it led to segfault.
We need to rethink proper fix but in the same time
cmd->mem pool is recreated per each lvm command so
this should not cause problems even when we run
multiple commands in lvm shell.
A valgrind snapshot of the corruption:
Invalid read of size 1
at 0x4C29F92: strlen (mc_replace_strmem.c:403)
by 0x5495F2E: dm_pool_strdup (pool.c:51)
by 0x1592A7: _create_id (pvmove.c:774)
by 0x159409: pvmove_poll (pvmove.c:796)
by 0x1599E3: pvmove (pvmove.c:931)
by 0x15105B: lvm_run_command (lvmcmdline.c:1655)
by 0x1523C3: lvm2_main (lvmcmdline.c:2121)
by 0x1754F3: main (lvm.c:22)
Address 0xf15df8a is 138 bytes inside a block of size 8,192 free'd
at 0x4C28430: free (vg_replace_malloc.c:446)
by 0x5494E73: dm_free_wrapper (dbg_malloc.c:357)
by 0x5495DE2: _free_chunk (pool-fast.c:318)
by 0x549561C: dm_pool_free (pool-fast.c:151)
by 0x164451: destroy_processing_handle (toollib.c:1837)
by 0x1598C1: pvmove (pvmove.c:903)
by 0x15105B: lvm_run_command (lvmcmdline.c:1655)
by 0x1523C3: lvm2_main (lvmcmdline.c:2121)
by 0x1754F3: main (lvm.c:22)
Fix regression caused by c9f021de0b.
This commit actually transfered real-action (e.g. device removal)
into the next loop which has however missed to check for break.
So add check for break also there.
When creating a list in 'context of command' - use proper mempool.
vg->vgmem is mempool related to VG metadata - and can be eventually
locked read-only when VG struct is shared.
The extent size must fits all blocks in 4294967295 sectors
(in 512b units) this is 1/2 KiB less then 2TiB.
So while previous statement 'suggested' 2TiB is still acceptable value,
make it clear it's not.
As now we support any multiples of 128KB as extent size -
values like 2047G will still 'flow-in' otherwise the largest power-of-2
supported value is 1TiB.
With 1TiB user needs 8388608 extents for 8EiB device.
(FYI such device is already unusable with todays glibc-2.22.90-27)
4GiB extent size is currently the smallest extent size which allows
a user to create 8EiB devices (with 2GiB it's less then 8EiB).
TODO: lvm2 may possibly print amount of 'lost/unused space' on a PV,
since using such ridiculously sized extent size may result in huge
space being left unaccessible.
Add a comment in _process_pvs_in_vg() to document the
place where there have been problems with processing
PVs twice.
For a while we had a hacky workaround here where we'd
skip processing a PV if its device wasn't found in
all_devices (and !is_missing_pv since we want to
process PVs with missing devices.). That workaround
was removed in commit 5cd4d46f because it was no
longer needed.
The workaround had originally been needed to prevent
a device from being processed twice when the PV had
no MDAs -- it would be processed once in its real VG
and then the workaround would prevent it from being
processed a second time in the orphan VG.
Wrongly appearing as an orphan likely happened because
lvmcache would consider the no-MDA PV an orphan unless
the real VG holding that PV was also in lvmcache.
This issue is also mentioned in pvchange where holding
the global lock allows VGs to remain in lvmcache so
PVs with 0 mdas are not considered orphans.
The workaround in _process_pvs_in_vg() was originally
intended for reporting commands, not for pvchange.
But, it was accidentally helping pvchange also because
the method described by the pvchange global lock
comment had been subverted by commit 80f4b4b8.
Commit 80f4b4b8 was found to be unnecessary, and was
reverted in commit e710bac0. This restored the
intended global lock lvmcache effect to pvchange, and
it no longer relied on the workaround in toollib.
Previously, pvmove used the function find_pv_in_vg() which did the
equivalent of process_each_pv() by doing:
find_pv_by_name() -> get_pvs() ->
get_pvs_internal() -> _get_pvs() -> get_vgids() ->
/* equivalent to process_each_pv */
dm_list_iterate_items(vgids)
vg = vg_read_internal()
dm_list_iterate_items(&vg->pvs)
With the found 'pv', it would do vg_read() on pv_vg_name(pv),
and then do the actual pvmove processing.
This commit simplifies by using process_each_pv() and putting
the actual pvmove processing into the "single" function.
This eliminates both find_pv_by_name() and the vg_read().
The processing code that followed vg_read remains the same.
The return code for the pvmove command is not based on the
process_each_pv return code, but is based on the success/fail
conditions in the existing code.
When an orphan PV is changed/resized, the
lvmlockd global lock is converted from sh
to ex. If the command is changing two
orphan PVs, the conversion to ex should
be done only once.
The problem addressed by this workaround no longer
seems to exist, so remove it. PVs with no mdas
no longer appear in both their actual VG and in
the orphan VG.
Use process_each_vg() to lock and read the old VG,
and then call the main vgrename code.
When real VG names are used (not a UUID in place of the
old name), the command still pre-locks the new name
(when strcmp wants it locked first), before calling
process_each_vg on the old name.
In the case where the old name is replaced with a UUID,
process_each_vg now translates that UUID into the real
VG name, which it locks and reads. In this case, we
cannot do pre-locking to maintain lock ordering because
the old name is unknown. So, in this case the strcmp
based lock ordering is suppressed and the old name is
always locked first. This opens a remote chance for
lock ordering conflict between racing vgrenames between
two names where one or both commands use the UUID.
Also always clear the internal lvmcache after rescanning, and
reinstate a test for --trustcache so that 'pvs --trustcache'
(for example) avoids rescanning.
