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.
Before 'archive()' is called, lvm2 must not touch/modify metadata.
So move setting CACHE_VOL related flags past this point.
Also make sure reading of cache segtype always restores this
flag properly (even if compatible flag would be lost).
When an LV is used as a writecache cachevol, give
it the LV name a _cvol suffix. Remove the suffix
when the cachevol is detached, restoring the
original LV name.
A cachevol LV had the CACHE_VOL status flag in metadata,
and the cache LV using it had no new flag. This caused
problems if the new metadata was used by an old version
of lvm. An old version of lvm would have two problems
processing the new metadata:
. The old lvm would return an error when reading the VG
metadata when it saw the unknown CACHE_VOL status flag.
. The old lvm would return an error when reading the VG
metadata because it would not find an expected cache pool
attached to the cache LV (since the cache LV had a
cachevol attached instead.)
Change the use of flags:
. Change the CACHE_VOL flag to be a COMPATIBLE flag (instead
of a STATUS flag) so that old versions will not fail when
they see it.
. When a cache LV is using a cachevol, the cache LV gets
a new SEGTYPE flag CACHE_USES_CACHEVOL. This flag is
appended to the segtype name, so that old lvm versions
will fail to use the LV because of an unknown segtype,
as opposed to failing to read the VG.
Instead of using 'noflush' option, switch cache_mode into WRITETHROUGH
which does not require flushing, when user confirmed he does not
want flushing for WRITEBACK (because of (partially) missing caching PV)
For wiping we activate and clear 'regular' devices,
since in case of whole process interuption (i.e. kill -9)
we leave metadata & DM table and workable state all the time.
vgck --updatemetadata would write the same correct
metadata to good mdas, and then to bad mdas, but the
sequence of vg_write/vg_commit calls betwen good and
bad mdas could cause a different description field to
be generated for good/bad mdas. (The description field
describing the command was recently included in the
ondisk copy of the metadata text.)
When the PV device names in the VG metadata do not match the
current PV device names seen on the system, do not use the
optimized activation function (that avoids extra device scanning.)
When the device names do not match, it's a clue that there could
be duplicate PVs, in which case we want to scan all devicess to
find any duplicates and stop the activation if found.
This does not prevent autoactivating a VG from the incorrect
duplicate PV, because the incorrect duplicate may appear by itself
first. At that point its duplicate PV does not exist to be seen.
(A future enhancement could use the WWID to strengthen this
detection.)
- use internal CACHE_VOL flag on cachevol LV
- add suffixes to dm uuids for internal LVs
- display appropriate letters in the LV attr field
- display writecache's cachevol in lvs output
. For dm-cache in writethrough, always allow splitcache,
whether the cache is missing PVs or not.
. For dm-cache in writeback, if the cache is missing PVs,
allow splitcache with force and yes.
. For dm-writecache, if the cache is missing PVs,
allow splitcache with force and yes.
Enhance 'activation' experience for VDO pool to more closely match
what happens for thin-pools where we do use a 'fake' LV to keep pool
running even when no thinLVs are active. This gives user a choice
whether he want to keep thin-pool running (wihout possibly lenghty
activation/deactivation process)
As we do plan to support multple VDO LVs to be mapped into a single VDO,
we want to give user same experience and 'use-patter' as with thin-pools.
This patch gives option to activate VDO pool only without activating
VDO LV.
Also due to 'fake' layering LV we can protect usage of VDO pool from
command like 'mkfs' which do require exlusive access to the volume,
which is no longer possible.
Note: VDO pool contains 1024 initial sectors as 'empty' header - such
header is also exposed in layered LV (as read-only LV).
For blkid we are indentified as LV with UUID suffix - thus private DM
device of lvm2 - so we do not need to store any extra info in this
header space (aka zero is good enough).
When pvscan is used to activate a VG via an
asynchronous service (i.e. lvm2-pvscan), there
is no requirement that the command wait for
udev to create device nodes before returning.
It's possible that waiting for udev is slow
enough to cause the service running the command
to time out. So, allow the --noudevsync option
to be given to pvscan to skip waiting for udev.
(This commit is not changing the lvm2-pvscan
service itself to use --noudevsync.)
Still unknown is whether there are any complex
LV activation cases in which lvm itself requires
access to a device node, in which case the udev
wait could be needed by lvm itself.
(When running an activation command directly
from the command line, it's generally expected
that the activated LVs are ready to use when
the command is finished, so lvm waits for
udev to finish creating the dev nodes.)
When an online PV completed a VG, the standard
activation functions were used to activate the VG.
These functions use a full scan of all devs.
When many pvscans are run during startup and need
to activate many VGs, scanning all devs from all
the pvscans can take a long time.
Optimize VG activation in pvscan to scan only the
devs in the VG being activated. This makes use of
the online file info that was used to determine
the VG was complete.
The downside of this approach is that pvscan activation
will not detect duplicate PVs and block activation,
where a normal activation command (which scans all
devices) would.
New udev in rawhide seems to be 'dropping' udev rule operations for devices
that are no longer existing - while this is 'probably' a bug - it's
revealing moments in lvm2 that likely should not run in a single
transaction and we should wait for a cookie before submitting more work.
TODO: it seem more 'error' paths should always include synchronization
before starting deactivating 'just activated' devices.
We should probably figure out some 'automatic' solution for this instead
of placing sync_local_dev_name() all over the place...
Between 'resume' and 'remove' we need to wait for udev to synchronize,
otherwise udev may 'skip' resume event processing if the udev node
is already gone.
Usually md components are eliminated in label scan and/or
duplicate resolution, but they could sometimes get into
the vg_read stage, where set_pv_devices compares the
device to the PV.
If set_pv_devices runs an md component check and finds
one, vg_read should eliminate the components.
In set_pv_devices, run an md component check always
if the PV is smaller than the device (this is not
very common.) If the PV is larger than the device,
(more common), do the component check when the config
setting is "auto" (the default).
Avoid having PVs with different logical block sizes in the same VG.
This prevents LVs from having mixed block sizes, which can produce
file system errors.
The new config setting devices/allow_mixed_block_sizes (default 0)
can be changed to 1 to return to the unrestricted mode.
An active md device with an end superblock causes lvm to
enable full md component detection. This was being done
within the filter loop instead of before, so the full
filtering of some devs could be missed.
Also incorporate the recently added config setting that
controls the md component detection.
Fix commit 7836e7aa1c
"pvscan: ignore device with incorrect size"
which caused pvscan to not consider a PV online (for purposes
of event based activation) if the PV and device sizes differed.
This helped to avoid mistaking MD components for PVs, and is
replaced by triggering an md component check when PV and device
sizes differ (which happens in set_pv_device).
This allows the creation of a striped mirror leg(s) during upconvert
by adding lvconvert command line options --stripes/--stripesize
for 'mirror' to tools/command-lines.in.
In case multiple mirror legs are being added, all will have the
same requested striped layout.
Resolves: rhbz1720705
The cache repair utility does not yet work with a cachevol
(where metadata and data exist on the same LV.) So, warn
and prompt if writeback is specified with a cachevol.
The exported VG checking/enforcement was scattered and
inconsistent. This centralizes it and makes it consistent,
following the existing approach for foreign and shared
VGs/PVs, which are very similar to exported VGs/PVs.
The access policy that now applies to foreign/shared/exported
VGs/PVs, is that if a foreign/shared/exported VG/PV is named
on the command line (i.e. explicitly requested by the user),
and the command is not permitted to operate on it because it
is foreign/shared/exported, then an access error is reported
and the command exits with an error. But, if the command is
processing all VGs/PVs, and happens to come across a
foreign/shared/exported VG/PV (that is not explicitly named on
the command line), then the command silently skips it and does
not produce an error.
A command using tags or --select handles inaccessible VGs/PVs
the same way as a command processing all VGs/PVs, and will
not report/return errors if these inaccessible VGs/PVs exist.
The new policy fixes the exit codes on a somewhat random set of
commands that previously exited with an error if they were
looking at all VGs/PVs and an exported VG existed on the system.
There should be no change to which commands are allowed/disallowed
on exported VGs/PVs.
Certain LV commands (lvs/lvdisplay/lvscan) would previously not
display LVs from an exported VG (for unknown reasons). This has
not changed. The lvm fullreport command would previously report
info about an exported VG but not about the LVs in it. This
has changed to include all info from the exported VG.
When vg_read rescans devices with the intention of
writing the VG, the label rescan can open the devs
RW so they do not need to be closed and reopened
RW in dev_write_bytes.
When monitoring, skip exported VGs without causing a command
failure.
The lvm2-monitor service runs 'vgchange --monitor y', so
any exported VG on the system would cause the service to
fail.
The man page generation for pvchange/lvchange/vgchange was
incorrect (leaving out some option listings) as a result of
commit e225bf5 "fix command definition for pvchange -a"
The -a was being included in the set of "one or more"
options instead of an actual required option. Even
though the cmd def was not implementing the restrictions
correctly, the command internally was.
Adjust the cmd def code which did not support a command
with some real required options and a set of "one or more"
options.
The way that this command now uses the global lock
followed by a label scan, it can simply check if the
new VG name exists, and if not lock it and create it.
These two flags may be not reset at the end of
the command when the unlock is implicit, which
is a problem if the cmd struct is reused.
Clear the flags in the general fin_locking.
The fact that vg repair is implemented as a part of vg read
has led to a messy and complicated implementation of vg_read,
and limited and uncontrolled repair capability. This splits
read and repair apart.
Summary
-------
- take all kinds of various repairs out of vg_read
- vg_read no longer writes anything
- vg_read now simply reads and returns vg metadata
- vg_read ignores bad or old copies of metadata
- vg_read proceeds with a single good copy of metadata
- improve error checks and handling when reading
- keep track of bad (corrupt) copies of metadata in lvmcache
- keep track of old (seqno) copies of metadata in lvmcache
- keep track of outdated PVs in lvmcache
- vg_write will do basic repairs
- new command vgck --updatemetdata will do all repairs
Details
-------
- In scan, do not delete dev from lvmcache if reading/processing fails;
the dev is still present, and removing it makes it look like the dev
is not there. Records are now kept about the problems with each PV
so they be fixed/repaired in the appropriate places.
- In scan, record a bad mda on failure, and delete the mda from
mda in use list so it will not be used by vg_read or vg_write,
only by repair.
- In scan, succeed if any good mda on a device is found, instead of
failing if any is bad. The bad/old copies of metadata should not
interfere with normal usage while good copies can be used.
- In scan, add a record of old mdas in lvmcache for later, do not repair
them while reading, and do not let them prevent us from finding and
using a good copy of metadata from elsewhere. One result is that
"inconsistent metadata" is no longer a read error, but instead a
record in lvmcache that can be addressed separate from the read.
- Treat a dev with no good mdas like a dev with no mdas, which is an
existing case we already handle.
- Don't use a fake vg "handle" for returning an error from vg_read,
or the vg_read_error function for getting that error number;
just return null if the vg cannot be read or used, and an error_flags
arg with flags set for the specific kind of error (which can be used
later for determining the kind of repair.)
- Saving an original copy of the vg metadata, for purposes of reverting
a write, is now done explicitly in vg_read instead of being hidden in
the vg_make_handle function.
- When a vg is not accessible due to "access restrictions" but is
otherwise fine, return the vg through the new error_vg arg so that
process_each_pv can skip the PVs in the VG while processing.
(This is a temporary accomodation for the way process_each_pv
tracks which devs have been looked at, and can be dropped later
when process_each_pv implementation dev tracking is changed.)
- vg_read does not try to fix or recover a vg, but now just reads the
metadata, checks access restrictions and returns it.
(Checking access restrictions might be better done outside of vg_read,
but this is a later improvement.)
- _vg_read now simply makes one attempt to read metadata from
each mda, and uses the most recent copy to return to the caller
in the form of a 'vg' struct.
(bad mdas were excluded during the scan and are not retried)
(old mdas were not excluded during scan and are retried here)
- vg_read uses _vg_read to get the latest copy of metadata from mdas,
and then makes various checks against it to produce warnings,
and to check if VG access is allowed (access restrictions include:
writable, foreign, shared, clustered, missing pvs).
- Things that were previously silently/automatically written by vg_read
that are now done by vg_write, based on the records made in lvmcache
during the scan and read:
. clearing the missing flag
. updating old copies of metadata
. clearing outdated pvs
. updating pv header flags
- Bad/corrupt metadata are now repaired; they were not before.
Test changes
------------
- A read command no longer writes the VG to repair it, so add a write
command to do a repair.
(inconsistent-metadata, unlost-pv)
- When a missing PV is removed from a VG, and then the device is
enabled again, vgck --updatemetadata is needed to clear the
outdated PV before it can be used again, where it wasn't before.
(lvconvert-repair-policy, lvconvert-repair-raid, lvconvert-repair,
mirror-vgreduce-removemissing, pv-ext-flags, unlost-pv)
Reading bad/old metadata
------------------------
- "bad metadata": the mda_header or metadata text has invalid fields
or can't be parsed by lvm. This is a form of corruption that would
not be caused by known failure scenarios. A checksum error is
typically included among the errors reported.
- "old metadata": a valid copy of the metadata that has a smaller seqno
than other copies of the metadata. This can happen if the device
failed, or io failed, or lvm failed while commiting new metadata
to all the metadata areas. Old metadata on a PV that has been
removed from the VG is the "outdated" case below.
When a VG has some PVs with bad/old metadata, lvm can simply ignore
the bad/old copies, and use a good copy. This is why there are
multiple copies of the metadata -- so it's available even when some
of the copies cannot be used. The bad/old copies do not have to be
repaired before the VG can be used (the repair can happen later.)
A PV with no good copies of the metadata simply falls back to being
treated like a PV with no mdas; a common and harmless configuration.
When bad/old metadata exists, lvm warns the user about it, and
suggests repairing it using a new metadata repair command.
Bad metadata in particular is something that users will want to
investigate and repair themselves, since it should not happen and
may indicate some other problem that needs to be fixed.
PVs with bad/old metadata are not the same as missing devices.
Missing devices will block various kinds of VG modification or
activation, but bad/old metadata will not.
Previously, lvm would attempt to repair bad/old metadata whenever
it was read. This was unnecessary since lvm does not require every
copy of the metadata to be used. It would also hide potential
problems that should be investigated by the user. It was also
dangerous in cases where the VG was on shared storage. The user
is now allowed to investigate potential problems and decide how
and when to repair them.
Repairing bad/old metadata
--------------------------
When label scan sees bad metadata in an mda, that mda is removed
from the lvmcache info->mdas list. This means that vg_read will
skip it, and not attempt to read/process it again. If it was
the only in-use mda on a PV, that PV is treated like a PV with
no mdas. It also means that vg_write will also skip the bad mda,
and not attempt to write new metadata to it. The only way to
repair bad metadata is with the metadata repair command.
When label scan sees old metadata in an mda, that mda is kept
in the lvmcache info->mdas list. This means that vg_read will
read/process it again, and likely see the same mismatch with
the other copies of the metadata. Like the label_scan, the
vg_read will simply ignore the old copy of the metadata and
use the latest copy. If the command is modifying the vg
(e.g. lvcreate), then vg_write, which writes new metadata to
every mda on info->mdas, will write the new metadata to the
mda that had the old version. If successful, this will resolve
the old metadata problem (without needing to run a metadata
repair command.)
Outdated PVs
------------
An outdated PV is a PV that has an old copy of VG metadata
that shows it is a member of the VG, but the latest copy of
the VG metadata does not include this PV. This happens if
the PV is disconnected, vgreduce --removemissing is run to
remove the PV from the VG, then the PV is reconnected.
In this case, the outdated PV needs have its outdated metadata
removed and the PV used flag needs to be cleared. This repair
will be done by the subsequent repair command. It is also done
if vgremove is run on the VG.
MISSING PVs
-----------
When a device is missing, most commands will refuse to modify
the VG. This is the simple case. More complicated is when
a command is allowed to modify the VG while it is missing a
device.
When a VG is written while a device is missing for one of it's PVs,
the VG metadata is written to disk with the MISSING flag on the PV
with the missing device. When the VG is next used, it is treated
as if the PV with the MISSING flag still has a missing device, even
if that device has reappeared.
If all LVs that were using a PV with the MISSING flag are removed
or repaired so that the MISSING PV is no longer used, then the
next time the VG metadata is written, the MISSING flag will be
dropped.
Alternative methods of clearing the MISSING flag are:
vgreduce --removemissing will remove PVs with missing devices,
or PVs with the MISSING flag where the device has reappeared.
vgextend --restoremissing will clear the MISSING flag on PVs
where the device has reappeared, allowing the VG to be used
normally. This must be done with caution since the reappeared
device may have old data that is inconsistent with data on other PVs.
Bad mda repair
--------------
The new command:
vgck --updatemetadata VG
first uses vg_write to repair old metadata, and other basic
issues mentioned above (old metadata, outdated PVs, pv_header
flags, MISSING_PV flags). It will also go further and repair
bad metadata:
. text metadata that has a bad checksum
. text metadata that is not parsable
. corrupt mda_header checksum and version fields
(To keep a clean diff, #if 0 is added around functions that
are replaced by new code. These commented functions are
removed by the following commit.)
uses vg_write to correct more common or less severe issues,
and also adds the ability to repair some metadata corruption
that couldn't be handled previously.
and implement it based on a device, not based
on a pv struct (which is not available when the
device is not a part of the vg.)
currently only the vgremove command wipes outdated
pvs until more advanced recovery is added in a
subsequent commit
If udev info is missing for a device, (which would indicate
if it's an MD component), then do an end-of-device read to
check if a PV is an MD component. (This is skipped when
using hints since we already know devs in hints are good.)
A new config setting md_component_checks can be used to
disable the additional end-of-device MD checks, or to
always enable end-of-device MD checks.
When both hints and udev info are disabled/unavailable,
the end of PVs will now be scanned by default. If md
devices with end-of-device superblocks are not being
used, the extra I/O overhead can be avoided by setting
md_component_checks="start".
Use the recently added dump routines to produce the
old/traditional pvck output, and remove the code that
had been used for that.
The validation/checking done by the new routines means
that new lines prefixed with CHECK are printed for
incorrect values.
Add 'pvck --dump headers' to print all the
lvm ondisk structs. Also checks the values
and prints any problems.
The previous dump metadata is also converted to
use these same routines, which do not depend on lvm
fully scanning/reading/processing the headers and
metadata on disk. This makes it useful to get data in
cases where there is corruption that would otherwise
prevent the normal functions from working.
The new command 'pvck --dump metadata PV' will extract
the current version of VG metadata from a PV for testing
and debugging. --dump metadata_area extracts the entire
text metadata area.
commit aa75b31db5
"pvscan: handle case of scanning PV without metadata last"
failed to recognize that an arg may be null in the case of
'pvscan --cache' (without -aay) which does not keep track
of complete VGs because it does not need to activate them.
The scanning rework missed removing this instance of label scan.
It's no longer needed because of the way that label scan is always
run once from the start of the command. This unnecessary scan
would be triggered by running 'pvs @tag'.
and don't call it from inside pvcreate_each_device.
This avoids having to repeat it for users of
pvcreate_each_device (pvcreate/pvremove/vgcreate/vgextend.)
There have been two file locks used to protect lvm
"global state": "ORPHANS" and "GLOBAL".
Commands that used the ORPHAN flock in exclusive mode:
pvcreate, pvremove, vgcreate, vgextend, vgremove,
vgcfgrestore
Commands that used the ORPHAN flock in shared mode:
vgimportclone, pvs, pvscan, pvresize, pvmove,
pvdisplay, pvchange, fullreport
Commands that used the GLOBAL flock in exclusive mode:
pvchange, pvscan, vgimportclone, vgscan
Commands that used the GLOBAL flock in shared mode:
pvscan --cache, pvs
The ORPHAN lock covers the important cases of serializing
the use of orphan PVs. It also partially covers the
reporting of orphan PVs (although not correctly as
explained below.)
The GLOBAL lock doesn't seem to have a clear purpose
(it may have eroded over time.)
Neither lock correctly protects the VG namespace, or
orphan PV properties.
To simplify and correct these issues, the two separate
flocks are combined into the one GLOBAL flock, and this flock
is used from the locking sites that are in place for the
lvmlockd global lock.
The logic behind the lvmlockd (distributed) global lock is
that any command that changes "global state" needs to take
the global lock in ex mode. Global state in lvm is: the list
of VG names, the set of orphan PVs, and any properties of
orphan PVs. Reading this global state can use the global lock
in sh mode to ensure it doesn't change while being reported.
The locking of global state now looks like:
lockd_global()
previously named lockd_gl(), acquires the distributed
global lock through lvmlockd. This is unchanged.
It serializes distributed lvm commands that are changing
global state. This is a no-op when lvmlockd is not in use.
lockf_global()
acquires an flock on a local file. It serializes local lvm
commands that are changing global state.
lock_global()
first calls lockf_global() to acquire the local flock for
global state, and if this succeeds, it calls lockd_global()
to acquire the distributed lock for global state.
Replace instances of lockd_gl() with lock_global(), so that the
existing sites for lvmlockd global state locking are now also
used for local file locking of global state. Remove the previous
file locking calls lock_vol(GLOBAL) and lock_vol(ORPHAN).
The following commands which change global state are now
serialized with the exclusive global flock:
pvchange (of orphan), pvresize (of orphan), pvcreate, pvremove,
vgcreate, vgextend, vgremove, vgreduce, vgrename,
vgcfgrestore, vgimportclone, vgmerge, vgsplit
Commands that use a shared flock to read global state (and will
be serialized against the prior list) are those that use
process_each functions that are based on processing a list of
all VG names, or all PVs. The list of all VGs or all PVs is
global state and the shared lock prevents those lists from
changing while the command is processing them.
The ORPHAN lock previously attempted to produce an accurate
listing of orphan PVs, but it was only acquired at the end of
the command during the fake vg_read of the fake orphan vg.
This is not when orphan PVs were determined; they were
determined by elimination beforehand by processing all real
VGs, and subtracting the PVs in the real VGs from the list
of all PVs that had been identified during the initial scan.
This is fixed by holding the single global lock in shared mode
while processing all VGs to determine the list of orphan PVs.
Handle the case where pvscan --cache -aay (with no dev args)
gets to the final PV, completing the VG, but that final PV does not
have VG metadata. In this case, we need to use VG metadata from a
previously scanned PV in the same VG, which we saved for this
possibility. Using this saved metadata, we can find which VG
this PVID belongs to, and then check if that VG is now complete,
and if so add the VG name to the list of complete VGs to be
autoactivated.
If a device looks like a PV, but its size does not
match the PV size in the metadata, then skip it for
purposes of autoactivation. It's probably not wrong
device for the PV.
In the past, the first 'pvscan --cache -aay dev' command
to run on the system would initialize the pvs_online dir
by scanning all devs and creating online files for all pvs
it found, and then autoactivating the VG (if complete) for
the named dev. The idea was that the system may not have
been able to run pvscan commands for early devices, so the
first pvscan to run would need to "make up" for any devices
that had appeared previously, which the system was unable to
scan. The problem or idea of making up for missed scans is
historical and should no longer be needed, so remove this
special init case.
When pvscan is run for the initialization case (the first
pvscan run on the system), it scans all devs and creates
online files for all PVs it finds. Previously it would
then autoactivate every complete VG, but change this to
only autoactive the (complete) VG corresponding to the
named device arg(s).
When lvextend extends an LV that is active with a shared
lock, use this as a signal that other hosts may also have
the LV active, with gfs2 mounted, and should have the LV
refreshed to reflect the new size. Use the libdlmcontrol
run api, which uses dlm_controld/corosync to run an
lvchange --refresh command on other cluster nodes.
Allow using caching with VDO.
User can either cache a single vdopool or
a vdo LV - difference when the caching is put-in depends on a use-case
and it's upto user to decide which kind of speed is expected.
Fix to previous commit
"pvscan: ignore online for shared and foreign PVs"
which was incorrectly considering a PV foreign if its
VG had no system ID when the host did have a system ID.
Activation would not be allowed anyway, but we can
check for these cases early and avoid wasted time in
pvscan managing online files an attempting activation.
and "cachepool" to refer to a cache on a cache pool object.
The problem was that the --cachepool option was being used
to refer to both a cache pool object, and to a standard LV
used for caching. This could be somewhat confusing, and it
made it less clear when each kind would be used. By
separating them, it's clear when a cachepool or a cachevol
should be used.
Previously:
- lvm would use the cache pool approach when the user passed
a cache-pool LV to the --cachepool option.
- lvm would use the cache vol approach when the user passed
a standard LV in the --cachepool option.
Now:
- lvm will always use the cache pool approach when the user
uses the --cachepool option.
- lvm will always use the cache vol approach when the user
uses the --cachevol option.
Without this, the output from different commands in a single
log file could not be separated.
Change the default "indent" setting to 0 so that the default
debug output does not include variable spaces in the middle
of debug lines.
When a VG has multiple PVs, and all those PVs come online
at the same time, concurrent pvscans for each PV will all
create the individual pvid files, and all will often see
the VG is now complete. This causes each of the pvscan
commands to think it should activate the VG, so there
are multiple activations of the same VG. The vg lock
serializes them, and only the first pvscan actually does
the activation, but there is still a lot of extra overhead
and time used by the other pvscans that attempt to
activate the already active VG. This can lead to a backlog
of pvscans and timeouts.
To fix this, this adds a new /run/lvm/vgs_online/ dir that
works like the existing /run/lvm/pvs_online/ dir. Each pvscan
that wants to activate a VG will first try to exlusively create
the file vgs_online/<vgname>. Only the first pvscan will
succeed, and that one will do the VG activation. The other
pvscans will find the vgname file exists and will not do the
activation step.
When a PV goes offline, the vgs_online file for the corresponding
VG is removed. This allows the VG to be autoactivated again
when the PV comes online again. This requires that the vgname be
stored in the pvid files.
Use a file lock to ensure that only one pvscan will do
initialization of pvs_online, otherwise multiple concurrent
pvscans may all see an empty pvs_online directory and
do initialization.
The pvscan that is doing initialization should also only
attempt to activate complete VGs.
When aay was included in the pvscan --cache command,
the activation part was complaining about the unusual
state of the hint file since it had been recreated
just prior.
Just like we support for thin-pool syntax:
lvcreate --thinpool new_tpoolname -L Size vg
add same support logic with for vdo-poo:
lvcreate --vdopool new_vpoolname -L Size vg
Also move description of syntax bellow thin-pool, so it's
correctly ordered in generated man page.
When using 'lvcreate -l100%VG' and there is big disproportion between
real available space and requested setting - automatically fallback
to 100%FREE.
Difference can be seen when VG is big and already most space was
allocated, so the requestion 100%VG can end (and by spec for % modifier
it's correct) as LV with size of 1%VG. Usually this is not a big
problem - buit in some cases - like cache-pool allocation, this
can result a big difference for chunksize selection.
With this patch it's more closely match common-sense logic without
the need of reitteration of too big changes in lvm2 core ATM.
TODO: in the future there should be allocator solving all allocations
in a single call.
An idea from Zdenek for better ensuring valid hints by invalidating
them when pvscan --cache <device> sees a new PV, which is a case
where we know that hints should be invalidated. This is triggered
from systemd/udev logic, and there may be some cases where it would
invalidate hints that the existing methods wouldn't detect.
If there are two independent scripts doing:
vgchange --lockstart vg
lvchange -ay vg/lv
The first vgchange to do the lockstart will wait for
the lockstart to complete before returning.
The second vgchange to do the lockstart will see that
the start is already in progress (from the first) and
will do nothing. This means the second does not wait
for any lockstart to complete, and moves on to the
lvchange which may find the lockspace still starting
and fail.
To fix this, make the vgchange lockstart command
wait for any lockstart's in progress to complete.
Save the list of PVs in /run/lvm/hints. These hints
are used to reduce scanning in a number of commands
to only the PVs on the system, or only the PVs in a
requested VG (rather than all devices on the system.)
When initializing an LV to hold the writecache, use wipe_lv()
which looks for specific signatures on the LV.
Wiping signatures is not necessary, but printing a warning
that names a specific signature (in addition to the existing
generic warning/confirmation) may help if a user accidentally
specifies the wrong LV which contains something important.
Detach function return 0 for error and 1 for success.
Add missing log errors from failing deactivation.
Add missing log error from failing synchronization.
Since configure.h is a generated header and it's missing traditional
ifdefs preambule - it can be included & parsed multiple times.
Normally compiler is fine when defines have same value and there is
no warning - yet we don't need to parse this several times
and by adding -include directive we can ensure every file
in the package is rightly compile with configure.h as the
first header file.
With older gcc - we need to resolve symbols linked with devmapper-event
that is now using -ldevmapper.
Also add forgotten systemd library needed for dbus notification.
Ensure configure.h is always 1st. included header.
Maybe we could eventually introduce gcc -include option, but for now
this better uses dependency tracking.
Also move _REENTRANT and _GNU_SOURCE into configure.h so it
doesn't need to be present in various source files.
This ensures consistent compilation of headers like stdio.h since
it may produce different declaration.
There's a small window during creation of a new RaidLV when
rmeta SubLVs are made visible to wipe them in order to prevent
erroneous discovery of stale RAID metadata. In case a crash
prevents the SubLVs from being committed hidden after such
wiping, the RaidLV can still be activated with the SubLVs visible.
During deactivation though, a deadlock occurs because the visible
SubLVs are deactivated before the RaidLV.
The patch adds _check_raid_sublvs to the raid validation in merge.c,
an activation check to activate.c (paranoid, because the merge.c check
will prevent activation in case of visible SubLVs) and shares the
existing wiping function _clear_lvs in raid_manip.c moved to lv_manip.c
and renamed to activate_and_wipe_lvlist to remove code duplication.
Whilst on it, introduce activate_and_wipe_lv to share with
(lvconvert|lvchange).c.
Resolves: rhbz1633167
In RHEL7 we marked mirrored mirror logs as deprecated and
added a related message. This patch prohibits creating new
'mirror' LVs with that log type or converting existing LVs
to have one.
Existing LVs with mirrored mirror log can be activated
and converted to disk/core logs.
Avoid double deprecation message when running lvconvert.
Resolves: rhbz1643562
. When using default settings, this commit should change
nothing. The first PE continues to be placed at 1 MiB
resulting in a metadata area size of 1020 KiB (for
4K page sizes; slightly smaller for larger page sizes.)
. When default_data_alignment is disabled in lvm.conf,
align pe_start at 1 MiB, based on a default metadata area
size that adapts to the page size. Previously, disabling
this option would result in mda_size that was too small
for common use, and produced a 64 KiB aligned pe_start.
. Customized pe_start and mda_size values continue to be
set as before in lvm.conf and command line.
. Remove the configure option for setting default_data_alignment
at build time.
. Improve alignment related option descriptions.
. Add section about alignment to pvcreate man page.
Previously, DEFAULT_PVMETADATASIZE was 255 sectors.
However, the fact that the config setting named
"default_data_alignment" has a default value of 1 (MiB)
meant that DEFAULT_PVMETADATASIZE was having no effect.
The metadata area size is the space between the start of
the metadata area (page size offset from the start of the
device) and the first PE (1 MiB by default due to
default_data_alignment 1.) The result is a 1020 KiB metadata
area on machines with 4KiB page size (1024 KiB - 4 KiB),
and smaller on machines with larger page size.
If default_data_alignment was set to 0 (disabled), then
DEFAULT_PVMETADATASIZE 255 would take effect, and produce a
metadata area that was 188 KiB and pe_start of 192 KiB.
This was too small for common use.
This is fixed by making the default metadata area size a
computed value that matches the value produced by
default_data_alignment.
instead of a separate --writecacheblocksize option.
writecache block_size is not technically a setting,
but it can borrow the option as a special case.
If a single, standard LV is specified as the cache, use
it directly instead of converting it into a cache-pool
object with two separate LVs (for data and metadata).
With a single LV as the cache, lvm will use blocks at the
beginning for metadata, and the rest for data. Separate
dm linear devices are set up to point at the metadata and
data areas of the LV. These dm devs are given to the
dm-cache target to use.
The single LV cache cannot be resized without recreating it.
If the --poolmetadata option is used to specify an LV for
metadata, then a cache pool will be created (with separate
LVs for data and metadata.)
Usage:
$ lvcreate -n main -L 128M vg /dev/loop0
$ lvcreate -n fast -L 64M vg /dev/loop1
$ lvs -a vg
LV VG Attr LSize Type Devices
main vg -wi-a----- 128.00m linear /dev/loop0(0)
fast vg -wi-a----- 64.00m linear /dev/loop1(0)
$ lvconvert --type cache --cachepool fast vg/main
$ lvs -a vg
LV VG Attr LSize Origin Pool Type Devices
[fast] vg Cwi---C--- 64.00m linear /dev/loop1(0)
main vg Cwi---C--- 128.00m [main_corig] [fast] cache main_corig(0)
[main_corig] vg owi---C--- 128.00m linear /dev/loop0(0)
$ lvchange -ay vg/main
$ dmsetup ls
vg-fast_cdata (253:4)
vg-fast_cmeta (253:5)
vg-main_corig (253:6)
vg-main (253:24)
vg-fast (253:3)
$ dmsetup table
vg-fast_cdata: 0 98304 linear 253:3 32768
vg-fast_cmeta: 0 32768 linear 253:3 0
vg-main_corig: 0 262144 linear 7:0 2048
vg-main: 0 262144 cache 253:5 253:4 253:6 128 2 metadata2 writethrough mq 0
vg-fast: 0 131072 linear 7:1 2048
$ lvchange -an vg/min
$ lvconvert --splitcache vg/main
$ lvs -a vg
LV VG Attr LSize Type Devices
fast vg -wi------- 64.00m linear /dev/loop1(0)
main vg -wi------- 128.00m linear /dev/loop0(0)
This fixes a problem in commit e6bb780d24, in which the
back compat handling for the old locking_type=4 was
incorrectly translated to mean the same thing as --readonly,
which prevented activation because activation uses an
exclusive vg lock. Previously, locking_type=4 allowed
activation.
If we see locking_type 4 in an old config, translate it to
the new combination of --readonly and --sysinit, which we
now define to mean the --readonly behavior with an exception
to allow activation.
When vgcreate does an automatic pvcreate, it opens the
dev with O_EXCL to ensure no other subsystem is using
the device. This exclusive fd remained in bcache and
prevented activation parts of lvm from using the dev.
This appeared with vgcreate of a sanlock VG because of
the unique combination where the dev is not yet a PV,
so pvcreate is needed, and the vgcreate also creates
and activates an internal LV for sanlock.
Fix this by closing the exclusive fd after it's used
by pvcreate so that it won't interfere with other
bits of lvm that may try to use the device.
Conversions of LVs under snapshot to thinpool or cachepool
correctly fail but leave them inactive and provide cryptic
error messages like 'Internal error: #LVs (10) != #visible
LVs (2) + #snapshots (1) + #internal LVs (5) in VG VG'.
Reject and provide better error message.
Resolves: rhbz1514146
The 'lvconvert LV' command def has caused multiple problems
for command matching because it matches the required options
of any lvconvert command. Any lvconvert with incorrect options
ends up matching 'lvconvert LV', which then produces an error
about incorrect options being used for 'lvconvert LV'. This
prevents suggestions from nearest-command partial command matches.
Add a special case for 'lvconvert LV' so that it won't be used
as a partial match for a command that has options specified.
Native disk scanning is now both reduced and
async/parallel, which makes it comparable in
performance (and often faster) when compared
to lvm using lvmetad.
Autoactivation now uses local temp files to record
online PVs, and no longer requires lvmetad.
There should be no apparent command-level change
in behavior.
When lvmetad is not used, use temporary files to record
which PVs have appeared. Use these temp files to determine
when a VG is complete, to trigger autoactivation.
This change allows us to remove lvmetad while keeping the
same autoactivation behavior that lvmetad provides.
The temp files are created in /run/lvm/pvs_online/ and are
named for the PVID of the PV. The files contain the
major:minor of the device the PV was read from.
e.g. if VG foo has dev1 and dev2, then:
. pvscan --cache -aay dev1
reads vg metadata from dev1
creates /run/lvm/pvs_online/<pvid-of-dev1>
checks if all vg->pvs are online: no
. pvscan --cache -aay dev2
reads vg metadata from dev2
creates /run/lvm/pvs_online/<pvid-of-dev2>
checks if all vg->pvs are online: yes
autoactivates vg
A 'pvscan --cache dev' (without -aay) still records that
dev is online.
A 'pvscan --cache --major X --minor Y' after a device is
gone will remove the temp file for it.
A 'pvscan --cache [-aay]' (no devs) resets the state of
temp files by removing them all, then scanning all devs
and creating temp files for PVs that are found.
If no online files exist, the first pvscan --cache scans
all devs and creates temp files for any PVs found.
The scope of the temp files is only pvscan, and they are only
used for pvscan-based autoactivation. No other commands are
concerned with or aware of these temp files. When lvm creates
or removes PVs, no attempt is made to update the temp files.
Support vgchange usage with VDO segtype.
Also changing extent size need small update for vdo virtual extent.
TODO: API needs enhancements so it's not about adding ifs() everywhere.
It's no longer needed. Clustered VGs are now handled in
the same way as foreign VGs, and as shared VGs that
can't be accessed:
- A command processing all VGs sees a clustered VG,
prints a message ("Skipping clustered VG foo."),
skips it, and does not fail.
- A command where the clustered VG is explicitly
named on the command line, prints a message and fails.
"Cannot access clustered VG foo, see lvmlockd(8)."
The option is listed in the set of ignored options for
the commands that previously accepted it. (Removing it
entirely would cause commands/scripts to fail if they
set it.)
Shuffle code for better readability as set of conditions was
hard to follow.
Make it obvious the refresh & activate path is handling
monitoring and polling on its own.
So the only --monitor and --poll option needs explicit care.
Option --monitor without option --poll will now as a result
of this patch NOT start polling.
So command: vgchange --monitor n is no longer a polling starter.
Restoring polling for activated volumes lost with my recent commit:
75fed05d3e and move start of polling
directly into _activate_lvs_in_vg() - as there we know exactly
if there was some volume even activated.
Also make it sharing same code for pvscan -aay.
The previous method for forcibly changing a clustered VG
to a local VG involved using -cn and locking_type 0.
Since those options are deprecated, replace it with
the same command used for other forced lock type changes:
vgchange --locktype none --lockopt force.
vgreduce, vgremove and vgcfgrestore were acquiring
the orphan lock in the midst of command processing
instead of at the start of the command. (The orphan
lock moved to being acquired at the start of the
command back when pvcreate/vgcreate/vgextend were
reworked based on pvcreate_each_device.)
vgsplit also needed a small update to avoid reacquiring
a VG lock that it already held (for the new VG name).
A few places were calling a function to check if a
VG lock was held. The only place it was actually
needed is for pvcreate which wants to do its own
locking (and scanning) around process_each_pv.
The locking/scanning exceptions for pvcreate in
process_each_pv/vg_read can be enabled by just passing
a couple of flags instead of checking if the VG is
already locked. This also means that these special
cases won't be enabled unknowingly in other places
where they shouldn't be used.
The last commit related to this was incomplete:
"Implement lock-override options without locking type"
This is further reworking and reduction of the locking.[ch]
layer which handled all clustering, but is now only used
for file locking. The "locking types" that this layer
implemented were removed previously, leaving only the
standard file locking. (Some cluster-related artifacts
remain to be cleared out later.)
Command options to override or modify locking behavior
are reimplemented here without using the locking types.
Also, deprecated locking_type values are recognized,
and implemented as if one of the equivalent override
options was set.
Options that override file locking are:
. --nolocking disables all file locking.
. --readonly grants read lock requests without actually
taking a file lock, and refuses write lock requests.
. --ignorelockingfailure tries to set up file locks and
uses them normally if possible. When not possible, it
behaves like --readonly, but allows activation.
. --sysinit is the same as ignorelockingfailure.
. global/metadata_read_only acquires actual read file
locks, and refuses write lock requests.
(Some of these options could probably be deprecated
because they were added as workarounds to various
locking_type behaviors that are now deprecated.)
The locking_type setting now has one valid value: 1 which
refers to standard file locking. Configs that contain
deprecated values are recognized and still work in
largely the same way:
. 0 disabled all locking, now implemented like --nolocking
is set. Allow the nolocking option in all commands.
. 1 is the normal file locking setting and is unchanged.
. 2 was for external locking which was not used, and
reverts to normal file locking.
. 3 was for cluster/clvm. This reverts to normal file
locking, and prints messages about lvmlockd.
. 4 was equivalent to readonly, now implemented like
--readonly is set.
. 5 disabled all locking, now implemented like
--nolocking is set.
The options: --nolocking, --readonly, --sysinit
override, or make exceptions to, the normal file locking
behavior. Implement these by just checking for the
options in the file locking path instead of using
special locking types.
Make activation commands:
vgchange -ay, lvchange -ay, pvscan -aay
take an exclusive file lock on the VG to serialize
multiple concurrent activation commands which could
otherwise interfere with each other.
Four commands lock two VGs at a time:
- vgsplit and vgmerge already have their own logic to
acquire the locks in the correct order.
- vgimportclone and vgrename disable this ordering check.
Different flavors of activate_lv() and lv_is_active()
which are meaningful in a clustered VG can be eliminated
and replaced with whatever that flavor already falls back
to in a local VG.
e.g. lv_is_active_exclusive_locally() is distinct from
lv_is_active() in a clustered VG, but in a local VG they
are equivalent. So, all instances of the variant are
replaced with the basic local equivalent.
For local VGs, the same behavior remains as before.
For shared VGs, lvmlockd was written with the explicit
requirement of local behavior from these functions
(lvmlockd requires locking_type 1), so the behavior
in shared VGs also remains the same.
"lvconvert --type {linear|striped|raid*} ..." on a striped/linear
LV provides convenience interim type to convert to the requested
final layout similar to the given raid* <-> raid* conveninece types.
Whilst on it, add missing raid5_n convenince type from raid5* to raid10.
Resolves: rhbz1439925
Resolves: rhbz1447809
Resolves: rhbz1573255
In this command, lvcreate creates a new LV and then combines
it with an existing cache pool, producing a cache LV. This
command was previously not allowed in in a shared VG.
When the lvmlockd lock is shared, upgrade it to ex
when repair (writing) is needed during vg_read.
Pass the lockd state through additional read-related
functions so the instances of repair scattered through
vg_read can be handled.
(Temporary solution until the ad hoc repairs can be
pulled out of vg_read into a top level, centralized
repair function.)
The report uses process_each_vg() which populates lvmcache
based on a VG list from lvmetad. If there are no VGs,
but only orphan PVs, the orphans are not shown. Add an
explicit call to populate lvmcache with PV info from lvmetad.
The device-mapper directory now holds a copy of libdm source. At
the moment this code is identical to libdm. Over time code will
migrate out to appropriate places (see doc/refactoring.txt).
The libdm directory still exists, and contains the source for the
libdevmapper shared library, which we will continue to ship (though
not neccessarily update).
All code using libdm should now use the version in device-mapper.
As we start refactoring the code to break dependencies (see doc/refactoring.txt),
I want us to use full paths in the includes (eg, #include "base/data-struct/list.h").
This makes it more obvious when we're breaking abstraction boundaries, eg, including a file in
metadata/ from base/
ATM it's a bit ugly to enforce flushing of 'stdio' here, but works as quick
hot-fix.
log_print*() is using buffered I/O.
But for pooling with typical 1s interval this may take a while before
buffer about continues progress gets flushed.
So ATM fflush().
TODO: either add log_print*_with_flush() or maybe directly use just
line buffering with log_print() and only log_debug() keep using buffered
I/O mode.
with the --labelsector option. We probably don't
need all this code to support any value for this
option; it's unclear how, when, why it would be
used.
Filters are still applied before any device reading or
the label scan, but any filter checks that want to read
the device are skipped and the device is flagged.
After bcache is populated, but before lvm looks for
devices (i.e. before label scan), the filters are
reapplied to the devices that were flagged above.
The filters will then find the data they need in
bcache.
The clvmd saved_vg data is independent from the normal lvm
lvmcache vginfo data, so separate saved_vg from vginfo.
Normal lvm doesn't need to use save_vg at all, and in clvmd,
lvmcache changes on vginfo can be made without worrying
about unwanted effects on saved_vg.
I don't like having this in a common header because it means you end
up including too much and causing unneccessary dependencies. eg,
lib/misc/lib.h includes libdevmapper.h, internationalisation, and
logging stuff.
There are likely more bits of code that can be removed,
e.g. lvm1/pool-specific bits of code that were identified
using FMT flags.
The vgconvert command can likely be reduced further.
The lvm1-specific config settings should probably have
some other fields set for proper deprecation.
Shift refresh of mirror table right into monitor_dev_for_events().
Use !vg_write_lock_held() to recognize use of lvchange/vgchange.
(this shall change if this would no longer work, but requires
futher some API changes).
With this patch dm mirror table is only refreshed when necassary.
Also update WARNING message about mirror usage without monitoring
and display LV name.
When pvmove was run in background mode and forks
instead of using lvmpolld, the child pvmove process
was not clearing the bcache from the parent, so all
the aio ops in the child were failing.
For reporting commands (pvs,vgs,lvs,pvdisplay,vgdisplay,lvdisplay)
we do not need to repeat the label scan of devices in vg_read if
they all had matching metadata in the initial label scan. The
data read by label scan can just be reused for the vg_read.
This cuts the amount of device i/o in half, from two reads of
each device to one. We have to be careful to avoid repairing
the VG if we've skipped rescanning. (The VG repair code is very
poor, and will be redone soon.)
Recent changes allow some major simplification of the way
lvmcache works and is used. lvmcache_label_scan is now
called in a controlled fashion at the start of commands,
and not via various unpredictable side effects. Remove
various calls to it from other places. lvmcache_label_scan
should not be called from anywhere during a command, because
it produces an incorrect representation of PVs with no MDAs,
and misclassifies them as orphans. This has been a long
standing problem. The invalid flag and rescanning based on
that is no longer used and removed. The 'force' variation is
no longer needed and removed.
When a PV is stacked on an LV, the LV will be kept in
bcache, and the open fd on the LV may interfere with
processing the LV. So, drop/close a bcache fd for
an LV before processing the LV.
Create a new dev->bcache_fd that the scanning code owns
and is in charge of opening/closing. This prevents other
parts of lvm code (which do various open/close) from
interfering with the bcache fd. A number of dev_open
and dev_close are removed from the reading path since
the read path now uses the bcache.
With that in place, open(O_EXCL) for pvcreate/pvremove
can then be fixed. That wouldn't work previously because
of other open fds.
In the same way as the other process_each functions.
In the common case all the info that's needed can be
used from lvmcache after a label scan. But this means
that unchosen devs for duplicate PVs need to be handled
explicitly.
The copy of VG metadata stored in lvmcache was not being used
in general. It pretended to be a generic VG metadata cache,
but was not being used except for clvmd activation. There
it was used to avoid reading from disk while devices were
suspended, i.e. in resume.
This removes the code that attempted to make this look
like a generic metadata cache, and replaces with with
something narrowly targetted to what it's actually used for.
This is a way of passing the VG from suspend to resume in
clvmd. Since in the case of clvmd one caller can't simply
pass the same VG to both suspend and resume, suspend needs
to stash the VG somewhere that resume can grab it from.
(resume doesn't want to read it from disk since devices
are suspended.) The lvmcache vginfo struct is used as a
convenient place to stash the VG to pass it from suspend
to resume, even though it isn't related to the lvmcache
or vginfo. These suspended_vg* vginfo fields should
not be used or touched anywhere else, they are only to
be used for passing the VG data from suspend to resume
in clvmd. The VG data being passed between suspend and
resume is never modified, and will only exist in the
brief period between suspend and resume in clvmd.
suspend has both old (current) and new (precommitted)
copies of the VG metadata. It stashes both of these in
the vginfo prior to suspending devices. When vg_commit
is successful, it sets a flag in vginfo as before,
signaling the transition from old to new metadata.
resume grabs the VG stashed by suspend. If the vg_commit
happened, it grabs the new VG, and if the vg_commit didn't
happen it grabs the old VG. The VG is then used to resume
LVs.
This isolates clvmd-specific code and usage from the
normal lvm vg_read code, making the code simpler and
the behavior easier to verify.
Sequence of operations:
- lv_suspend() has both vg_old and vg_new
and stashes a copy of each onto the vginfo:
lvmcache_save_suspended_vg(vg_old);
lvmcache_save_suspended_vg(vg_new);
- vg_commit() happens, which causes all clvmd
instances to call lvmcache_commit_metadata(vg).
A flag is set in the vginfo indicating the
transition from the old to new VG:
vginfo->suspended_vg_committed = 1;
- lv_resume() needs either vg_old or vg_new
to use in resuming LVs. It doesn't want to
read the VG from disk since devices are
suspended, so it gets the VG stashed by
lv_suspend:
vg = lvmcache_get_suspended_vg(vgid);
If the vg_commit did not happen, suspended_vg_committed
will not be set, and in this case, lvmcache_get_suspended_vg()
will return the old VG instead of the new VG, and it will
resume LVs based on the old metadata.
The old code was doing unnecessary label scans when
checking to see if the new VG name exists. A single
label_scan is sufficient if it is done after the
new VG lock is held.
