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.
It's possible for a dev-cache entry to remain after all
paths for it have been removed, and other parts of the
code expect that a dev always has a name. A better fix
may be to remove a device from dev-cache after all paths
to it have been removed.
When either logical block size or physical block size is 4K,
then lvmlockd creates sanlock leases based on 4K sectors,
but the lvm client side would create the internal lvmlock LV
based on the first logical block size it saw in the VG,
which could be 512. This could cause the lvmlock LV to be
too small to hold all the sanlock leases. Make the lvm client
side use the same sizing logic as lvmlockd.
dm-integrity stores checksums of the data written to an
LV, and returns an error if data read from the LV does
not match the previously saved checksum. When used on
raid images, dm-raid will correct the error by reading
the block from another image, and the device user sees
no error. The integrity metadata (checksums) are stored
on an internal LV allocated by lvm for each linear image.
The internal LV is allocated on the same PV as the image.
Create a raid LV with an integrity layer over each
raid image (for raid levels 1,4,5,6,10):
lvcreate --type raidN --raidintegrity y [options]
Add an integrity layer to images of an existing raid LV:
lvconvert --raidintegrity y LV
Remove the integrity layer from images of a raid LV:
lvconvert --raidintegrity n LV
Settings
Use --raidintegritymode journal|bitmap (journal is default)
to configure the method used by dm-integrity to ensure
crash consistency.
Initialization
When integrity is added to an LV, the kernel needs to
initialize the integrity metadata/checksums for all blocks
in the LV. The data corruption checking performed by
dm-integrity will only operate on areas of the LV that
are already initialized. The progress of integrity
initialization is reported by the "syncpercent" LV
reporting field (and under the Cpy%Sync lvs column.)
Example: create a raid1 LV with integrity:
$ lvcreate --type raid1 -m1 --raidintegrity y -n rr -L1G foo
Creating integrity metadata LV rr_rimage_0_imeta with size 12.00 MiB.
Logical volume "rr_rimage_0_imeta" created.
Creating integrity metadata LV rr_rimage_1_imeta with size 12.00 MiB.
Logical volume "rr_rimage_1_imeta" created.
Logical volume "rr" created.
$ lvs -a foo
LV VG Attr LSize Origin Cpy%Sync
rr foo rwi-a-r--- 1.00g 4.93
[rr_rimage_0] foo gwi-aor--- 1.00g [rr_rimage_0_iorig] 41.02
[rr_rimage_0_imeta] foo ewi-ao---- 12.00m
[rr_rimage_0_iorig] foo -wi-ao---- 1.00g
[rr_rimage_1] foo gwi-aor--- 1.00g [rr_rimage_1_iorig] 39.45
[rr_rimage_1_imeta] foo ewi-ao---- 12.00m
[rr_rimage_1_iorig] foo -wi-ao---- 1.00g
[rr_rmeta_0] foo ewi-aor--- 4.00m
[rr_rmeta_1] foo ewi-aor--- 4.00m
When vdopool is activated standalone - we use a wrapping linear device
to hold actual vdo device active - for this we can set-up read-only
device to ensure there cannot be made write through this device to
actual pool device.
Creating a snapshot was using a persistent LV lock
on the origin, so if the origin LV was inactive at
the time of the snapshot the LV lock would remain.
(Running lvchange -an on the inactive LV would
clear the LV lock.) Use a transient LV lock so it
will be dropped if it was not locked previously.
When formating VDO volume, the calculated amound of bits
for 'vdoformat --slab-bits' parameter was shifted by 2 bits
(calculated size was making 2MiB vdo_slab_size_mb value appear like if
user would be specifying only 512KiB)
Fixed by properly converting internal size_mb value to KiB.
Fix the anoying kernel message reported:
device-mapper: cache: 253:2: metadata operation 'dm_cache_commit' failed: error = -5
which has been reported while cachevol has been removed.
Happened via confusing variable - so switch the variable to commonly user '_size'
which presents a value in sector units and avoid 'scaling' this as extent length
by vg extent size when placing 'error' target on removal path.
Patch shouldn't have impact on actual users data, since at this moment
of removal all date should have been already flushed to origin device.
m
The previous patch improved read of pipe when lvm2 was looking
for default logical size, but we clearly must read pipe also
for -V case, when the logical size is already defined.
Still the place can be better to block only particular reshape
operations which ATM cause kernel problems.
We check if the new number of images is higher - and prevent to take
conversion if the volume is in use (i.e. thin-pool's data LV).
clang: it's supposedly impossible path to hit, as we should always
have origin_lv defined when running this path, but adding protection
isn't a big issue to make this obvious to analyzer.
Since _reserve_area() may fail due to error allocation failure,
add support to report this already reported failure upward.
FIXME: it's log_error() without causing direct command failure.
Although we expect min_chunk_size to be 32bit value, for
large size of caches it might be useful to do calcs 64bit.
So to avoid doing shift as signed 32bit - use unsigned 64bit
from the start.
reporting fields (-o) directly from kernel:
writecache_total_blocks
writecache_free_blocks
writecache_writeback_blocks
writecache_error
The data_percent field shows used cache blocks / total cache blocks.
Until we resolve reshape for 'stacked' devices, we need to disable it.
So users can no longer reshape i.e. thin-pool data volumes, causing
ATM bad thin-pool problems.
After the VG lock is taken for vg_read, reread the mda_header
and compare the metadata text offset and checksum to what was
seen during label scan. If it is unchanged, then the metadata
has not changed since the label scan, and the metadata does not
need to be reread under the lock for command processing.
For commands that do not make changes (e.g. reporting), the
mda_header is reread and checked on one mda to decide if the
full metadata rereading can be skipped. For other commands
(e.g. modifying the vg) the mda_header is reread and checked
from all PVs. (These could probably just check one mda also.)
When pvcreate/pvremove prompt the user, they first release
the global lock, then acquire it again after the prompt,
to avoid blocking other commands while waiting for a user
response. This release/reacquire changes the locking
order with respect to the hints flock (and potentially other
locks). So, to avoid deadlock, use a nonblocking request
when reacquiring the global lock.
Avoid mem leaking hint on every loop continue and
allocate hint only when it's going to be added into list.
Switch to use 'dm_strncpy()' and validate sizes.
For dev_in_device_list() != 0 allocated 'devl' was
actually leaking - so instead allocate 'devl' only
when !dev_in_device_list() and indent code around.
Since we check for NULL pointers earlier we need
to be consistent across function - since the NULL
would applies across whole function.
When dropping 'mda' check - we are actually
already dereferencing it before - so it can't
be NULL at that places (and it's validated
before entering _read_mda_header_and_metadata).
dev_unset_last_byte() must be called while the fd is still valid.
After a write error, dev_unset_last_byte() must be called before
closing the dev and resetting the fd.
In the write error path, dev_unset_last_byte() was being called
after label_scan_invalidate() which meant that it would not unset
the last_byte values.
After a write error, dev_unset_last_byte() is now called in
dev_write_bytes() before label_scan_invalidate(), instead of by
the caller of dev_write_bytes().
In the common case of a successful write, the sequence is still:
dev_set_last_byte(); dev_write_bytes(); dev_unset_last_byte();
Signed-off-by: Zhao Heming <heming.zhao@suse.com>
When resizing 2 volumes like thin-pool and it's metadata and they
would be of a different type - command would be actually expecting
both LVs being of a same segtype - and would throw an error in
case they are different.
This patch fixes is by setting a new segtype from last segment of
2nd. extented device.
Also it fixes the possible 'percentage' extension setup that
might have been used for 'primary' volume - while the 'secondary'
LV always goes with direct size - as we do not support 'percentage'
setup for them
This affects maily usage of thin-pool where the extension of
thin-pool data size may also lead to extension of metadata size.
Instead of checking all LVs in a VG - do just a direct copy of LVs
from the existing list ->segs_using_thin_lv.
TODO: maybe it could be better to expose seg_list to /tools...
Enhance lv_info with lv_info_with_name_check.
This 'variant' not only check existance if UUID in DM table
but also compares its DM name whether it's matching expected LV name.
Otherwise activation may 'skip' activation with rename in case the
DM UUID already exists, just device is different name.
This change make fairly easier manipulation with i.e. detached mirror
leg which ATM is using same UUID - just the LV name have been changed.
Used code was not able to run 'activation' (and do a rename) and just
skipped the call. So the code used to do a workaround and 'tried'
to deactivate such LV firts - this however work only in non-clvmd case,
as cluster was not having the lock for deactivated LV.
With this extended lv_info code will run 'activation' and will
synchronize the name to match expected LV name.
Patch extends _lv_info() with new paramter 'with_name_check',
which is later translated into 'name_check' argument for
_info_run() which in case of name mismatch evaluates the
check as if device does not exists.
Such call is only used in one place _lv_activate() which then
let activation run. All other invocation of _info() calls
are left intact.
TODO: fix mirror table manipulation (and raid)....
The return value from bcache_invalidate_fd() was not being checked.
So I've introduced a little function, _invalidate_fd() that always
calls bcache_abort_fd() if the write fails.
The resume of 'released' 'COW' should preceed the resume of origin.
The fact we need to do the sequence differently for merge was
cause by bugs fixed in 2 previous commits - so we no longer need
to recognize 'merging' and we should always go with single
sequence.
The importance of this order is - to properly remove '-real' device
from origin LV. When COW is activated as 2nd. '-real' device is
kept in table as it cannot be removed during 1st. resume of origin,
and later activation of COW LV no longer builds tree associated
with origin LV.
When checking device id of a thin device that is just being
merged - the snapshot actually could have been already finished
which means '-real' suffix for the LV is already gone and just LV
is there - so check explicitely for this condition and use
correct UUID for this case.
When a cachevol LV is attached, have the LV keep it's lock
allocated. The lock on the cachevol won't be used while
it's attached. When the cachevol is split a new lock does
not need to be allocated. (Applies to cachevol usage by
both dm-cache and dm-writecache.)
When LV gets cached and uses cache-pool - such cache-pool
will now get _cpool suffix automatically.
Thus 'Pool' column for cached LV will now show either _cvol
or _cpool LV.
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).
Since code is using -cdata and -cmeta UUID suffixes, it does not need
any new 'extra' ID to be generated and stored in metadata.
Since introduce of new 'segtype' cache+CACHE_USES_CACHEVOL we can
safely assume 'new' cache with cachevol will now be created
without extra metadata_id and data_id in metadata.
For backward compatibility, code still reads them in case older
version of metadata have them - so it still should be able
to activate such volumes.
Bonus is lowered size of lv structure used to store info about LV
(noticable with big volume groups).
The first part of a cachevol LV is used for metadata,
and the rest of the space is used for data. The
division of space between metadata and data depends
on the total size of the cachevol.
The previous division gave more space than needed to
metadata, it was:
cachevol size 8M to 128M -> metadata size 16M *
cachevol size 128M to 1G -> metadata size 32M
cachevol size 1G and up -> metadata size 64M
(* if this resulted in over half the LV used as
metadata, then half the cachevol would be used
for metadata, and the other half for data.)
The division of space now gives less space to
metadata, it is:
cachevol size 8M to 16M -> metadata size 4M
cachevol size 16M to 4G -> metadata size 8M
cachevol size 4G to 16G -> metadata size 16M
cachevol size 16G to 32G -> metadata size 32M
cachevol size 32G and up -> metadata size 64M
When a VG contains some LVs with unknown segtypes, the user
should still be allowed to activate other LVs in the VG that
are understood.
$ lvs foo
WARNING: Unrecognised flag CACHE_USES_CACHEVOL in segment type cache+CACHE_USES_CACHEVOL.
WARNING: Unrecognised segment type cache+CACHE_USES_CACHEVOL
LV VG Attr LSize
lvol0 foo -wi------- 4.00m
other foo vwi---u--- 48.00m
$ lvcreate -l1 foo
WARNING: Unrecognised flag CACHE_USES_CACHEVOL in segment type cache+CACHE_USES_CACHEVOL.
WARNING: Unrecognised segment type cache+CACHE_USES_CACHEVOL
Cannot change VG foo with unknown segments in it!
Cannot process volume group foo
$ lvchange -ay foo/lvol0
WARNING: Unrecognised flag CACHE_USES_CACHEVOL in segment type cache+CACHE_USES_CACHEVOL.
WARNING: Unrecognised segment type cache+CACHE_USES_CACHEVOL
$ lvchange -ay foo/other
WARNING: Unrecognised flag CACHE_USES_CACHEVOL in segment type cache+CACHE_USES_CACHEVOL.
WARNING: Unrecognised segment type cache+CACHE_USES_CACHEVOL
Refusing activation of LV foo/other containing an unrecognised segment.
$ lvs foo
WARNING: Unrecognised flag CACHE_USES_CACHEVOL in segment type cache+CACHE_USES_CACHEVOL.
WARNING: Unrecognised segment type cache+CACHE_USES_CACHEVOL
LV VG Attr LSize
lvol0 foo -wi-a----- 4.00m
other foo vwi---u--- 48.00m
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.
Enhance activation of cached devices using cachevol.
Correctly instatiace cachevol -cdata & -cmeta devices with
'-' in name (as they are only layered devices).
Code is also a bit more compacted (although still not ideal,
as the usage of extra UUIDs stored in metadata is troublesome
and will be repaired later).
NOTE: this patch my brink potentially minor incompatiblity for 'runtime' upgrade
Let vgck --updatemetadata repair cases where different mdas
hold indepedently valid but unmatching copies of the metadata,
i.e. different text metadata checksums or text metadata sizes.
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.)
If a VG is forcibly changed from lock_type sanlock to
lock_type none, the internal lvmlock LV is left behind.
If that LV is not removed before vgremove is run on the
VG, then an internal check will be triggered by the
hidden lvmlock LV. So, check for and remove a left over
lvmlock LV during vgremove.
Add lots of vdo fields:
vdo_operating_mode - For vdo pools, its current operating mode.
vdo_compression_state - For vdo pools, whether compression is running.
vdo_index_state - For vdo pools, state of index for deduplication.
vdo_used_size - For vdo pools, currently used space.
vdo_saving_percent - For vdo pools, percentage of saved space.
vdo_compression - Set for compressed LV (vdopool).
vdo_deduplication - Set for deduplicated LV (vdopool).
vdo_use_metadata_hints - Use REQ_SYNC for writes (vdopool).
vdo_minimum_io_size - Minimum acceptable IO size (vdopool).
vdo_block_map_cache_size - Allocated caching size (vdopool).
vdo_block_map_era_length - Speed of cache writes (vdopool).
vdo_use_sparse_index - Sparse indexing (vdopool).
vdo_index_memory_size - Allocated indexing memory (vdopool).
vdo_slab_size - Increment size for growing (vdopool).
vdo_ack_threads - Acknowledging threads (vdopool).
vdo_bio_threads - IO submitting threads (vdopool).
vdo_bio_rotation - IO enqueue (vdopool).
vdo_cpu_threads - CPU threads for compression and hashing (vdopool).
vdo_hash_zone_threads - Threads for subdivide parts (vdopool).
vdo_logical_threads - Logical threads for subdivide parts (vdopool).
vdo_physical_threads - Physical threads for subdivide parts (vdopool).
vdo_max_discard - Maximum discard size volume can recieve (vdopool).
vdo_write_policy - Specified write policy (vdopool).
vdo_header_size - Header size at front of vdopool.
Previously only 'lvdisplay -m' was exposing them.
Since we now support activation of 'vdo' volume
without explicit activation of 'vdopool' it's now possible
to have active layer vdopool (-vpool) volume and
having vdopool itself inactive - yet still in this
case we can show available stats for this volume.
But we need to show correct activation status and other
standard info.
Avoid checking 'lv_is_active()' since special LV types does this
validation anyway what calling _percent() function and call it
ONLY when none of special types is queried.
This restores support for VDO resize (as with support for
separate VDO pool activation, plain query for lv_is_active()
is not working in this case).
If the linear mapping is lost (for whatever reason, i.e.
test suite forcible 'dmsetup remove' linear LV,
lvm2 had hard times figuring out how to deactivate such DM table.
So add function which is in case inactive VDO pool LV checks if
the pool is actually still active (-vpool device present) and
it has open count == 0. In this case deactivation is allowed
to continue and cleanup DM table.
- 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 lvm2 is activating layered pool LV (to basically keep pool opened,
the other function used to be 'locking' be in sync with DM table)
use this LV in read-only mode - this prevents 'write' access into
data volume content of thin-pool.
Note: since EMPTY/unused thin-pool is created as 'public LV' for generic
use by any user who i.e. wish to maintain thin-pool and thins himself.
At this moment, thin-pool appears as writable LV. As soon as the 1st.
thinLV is created, layer volume will appear is 'read-only' LV from this moment.
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.
Fixes a regression from commit ba7ff96faf
"improve reading and repairing vg metadata"
where the error path for a vg name with invalid
charaters was missing an error flag, which led
to the caller not recognizing an error occured.
Previously, an error flag was hidden in the old
_vg_make_handle function.
Only the first entry of the filter array was being
included in the copy of the filter, rather than the
entire thing. The result is that hints would not be
refreshed if the filter was changed but the first
entry was unchanged.
Fixes a segfault in the recent commit e01fddc57:
"improve duplicate pv handling for md components"
While choosing between duplicates, the info struct is
not always valid; it may have been dropped already.
Remove the code that was still using the info struct for
size comparisons. The size comparisons were a bogus check
anyway because it was just preferring the dev that had
already been chosen, it wasn't actually comparing the
dev size to the PV size. It would be good to use a
dev/PV size comparison in the duplicate handling code, but
the PV size is not available until after vg_read, not
from the scan.
Since we need to preserve allocated strings across 2 separate
activation calls of '_tree_action()' we need to use other mem
pool them dm->mem - but since cmd->mem is released between
individual lvm2 locking calls, we rather introduce a new separate
mem pool just for pending deletes with easy to see life-span.
(not using 'libmem' as it would basicaly keep allocations over
the whole lifetime of clvmd)
This patch is fixing previous commmit where the memory was
improperly used after pool release.
Update configure and make code compilable if prlimit() is not present.
Since the code is suspicious do not cope yet with it's replacement
with set/getrlimit().
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...
Support internal removal of 'cache origin' volume - which we
do not normally expose to a user - however internal processing
loops may hit this condition (depending on order of list LVs).
So when this operation is internally requested - we automatically
try to remove it's 'holding' LV (cache LV) - which will also
remove the origin.
Drop the 'cluster-only' optimization so we do resume ALL device
before we try to wait on cookie before 'removal' operation.
It's more correct order of operation - alhtough possibly slightly
less efficient - but until we have correct list of operations
'in-progress' we can't do anything better.
With previous patch 30a98e4d67 we
started to put devices one pending_delete list instead
of directly scheduling their removal.
However we have operations like 'snapshot merge' where we are
resuming device tree in 2 subsequent activation calls - so
1st such call will still have suspened devices and no chance
to push 'remove' ioctl.
Since we curently cannot easily solve this by doing just single
activation call (which would be preferred solution) - we introduce
a preservation of pending_delete via command structure and
then restore it on next activation call.
This way we keep to remove devices later - although it might be
not the best moment - this may need futher tunning.
Also we don't keep the list of operation in 1 trasaction
(unless we do verify udev symlinks) - this could probably
also make it more correct in terms of which 'remove' can
be combined we already running 'resume'.
Resuming of 'error' table entry followed with it's dirrect removal
is now troublesame with latest udev as it may skip processing of
udev rules for already 'dropped' device nodes.
As we cannot 'synchronize' with udev while we know we have devices
in suspended state - rework 'cleanup' so it collects nodes
for removal into pending_delete list and process the list with
synchronization once we are without any suspended nodes.
When pvmove is finished, we do a tricky operation since we try to
resume multiple different device that were all joined into 1 big tree.
Currently we use the infromation from existing live DM table,
where we can get list of all holders of pvmove device.
We look for these nodes (by uuid) in new metadata, and we do now a full
regular device add into dm tree structure. All devices should be
already PRELOAD with correct table before entering suspend state,
however for correctly working readahead we need to put correct info
also into RESUME tree. Since table are preloaded, the same table
is skip and resume, but correct read ahead is now set.
Eliminate md components at the start so they don't
interfere with actual duplicates, and don't need
to be removed later. This also allows for choosing
no copy of a PVID if they all happen to be md
components.
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).
When there are more devices than the current soft
open file limit (default 1024), raise the soft limit
to the hard/max limit (default 4096).
Do this prior to scanning in case enough of the
devices are PVs that need to be kept open.
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.
Do this at two levels, although one would be enough to
fix the problem seen recently:
- Ignore any reported sector size other than 512 of 4096.
If either sector size (physical or logical) is reported
as 512, then use 512. If neither are reported as 512,
and one or the other is reported as 4096, then use 4096.
If neither is reported as either 512 or 4096, then use 512.
- When rounding up a limited write in bcache to be a multiple
of the sector size, check that the resulting write size is
not larger than the bcache block itself. (This shouldn't
happen if the sector size is 512 or 4096.)
Previously, consecutive copies of metadata would have garbage
data in the space between them. After metadata wrapping,
the garbage would be portions of old metadata. This made
analysis of the metadata area more difficult.
This would happen because the start of new copy of metadata
is advanced from the end of the last copy to start at the
next 512 byte boundary.
Zero the space between consecutive copies of metadata by
extending each metadata write to end at the next 512 byte
boundary. The size of the metadata itself is not extended,
only the write. The buffer being written contains the
metadata text followed by the necessary number of zeros.
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.
This check was mistakenly removed when shifting code in commit
"separate code for setting devices from metadata parsing".
Put it back with some new conditions.
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.
Previously the VG metadata description field (which contains
the command line) was only included in backup/archive copies
of the metadata. Now also include it in the metadata written
to the metadata areas.
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
The vg read and vg write cases need to update lvmcache
differently, so create separate functions for them.
The read case now handles checking for outdated mdas
and moves them aside into a new list to be repaired in
a subsequent commit.
The existing comment was desribing the correct behavior,
but the code didn't match. The commit is successful if
one mda was committed. Making it depend on the result of
the internal lvmcache update was wrong.
Have the caller pass the label_sector to the read
function so the read function can set the sector
field in the label struct, instead of having the
read function return a pointer to the label for
the caller to set the sector field.
Also have the read function return a flag indicating
to the caller that the scanned device was identified
as a duplicate pv.
Outdated PVs hold metadata for VG from which they
have been removed. Add the ability to keep track
of these in lvmcache.
This will be used for more advanced repair in a
subsequent commit.
mda's that cannot be processed by lvm because of
some corruption can be kept on a separate list.
These will be used for more advanced repair in a
subsequent commit.
When reading metadata headers and text, use a new set
of flags to identify specific errors that are seen.
These will be used for more advanced repair 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".
Save the previous duplicate PVs in a global list instead
of a list on the cmd struct. dmeventd reuses the cmd struct
for multiple commands, and the list entries between commands
were being freed (apparently), causing a segfault in dmeventd
when it tried to use items in cmd->unused_duplicate_devs
that had been saved there by the previous command.
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.
Recent kernel version from kernel commit:
de7180ff908b2bc0342e832dbdaa9a5f1ecaa33a
started to report in cache status line new flag:
no_discard_passdown
Whenever lvm spots unknown status it reports:
Unknown feature in status:
So add reconginzing this feature flag and also report this with
'lvs -o+kernel_discards'
When no_discard_passdown is found in status 'nopassdown' gets reported
for this field (roughly matching what we report for thin-pools).
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.
If an md component is not excluded by other means and
vg_read is used to read metadata from it, then this new
check compares the device size with the PV size, and runs
a full md check on the device if the sizes don't match.
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.
wipe_lv knows it's going to write the device, so it
can open rw from the start. It was opening readonly,
and then dev_write needed to reopen it readwrite.
When hints are invalid and ignored, the list of hints
could be non-empty (from additions before an invalid
hint was found). This confused the calling code which
was checking for an empty list to see if hints were used.
Ensure the list is empty when hints are not used.
This reverts 518a8e8cfb
"lvmlockd: activate mirror LVs in shared mode with cmirrord"
because while activating a mirror LV with cmirrord worked,
changes to the active cmirror did not work.
When data are growing, adapt also size of metadata.
As we get way too many reports from users doing huge growths of
data portion while keep metadata small and avoiding using monitoring.
So to enhance the user-experience in case user requests grown of
thin-pool (without passing PV list for growth) - lvm2 will automaticaly
grown also the metadata part of thin-pool (if possible).
Add function for estimation of thin-pool metadata size for given size of
data. Function is using already existing internal API so it can
be reused for resize of thin-pool data.
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.
When an LV is active with a shared lock, a command can be
run to change the LV with --lockopt skiplv (to override the
exclusive lock the command ordinarily requires which is not
compatible with the outstanding shared lock.)
In this case, other commands may have the LV active and may
need to refresh the LV, so print warning stating this.
Udev is running udev-rule action upon 'resume'.
However lvm2 in special case is doing replacement of
'soon-to-be-removed' device with 'error' target for resuming
and then follows actual removal - the sequence is usually quick,
so when udev start action - it can result in 'strange' error
message in kernel log like:
Process '/usr/sbin/dmsetup info -j 253 -m 17 -c --nameprefixes --noheadings --rows -o name,uuid,suspended' failed with exit code 1.
To avoid this - we need to ensure there is synchronization wait for udev
between 'resume' and 'remove' part of this process.
However existing code put strict requirement to avoid synchronizing with
udev inside critical section - but this originally came from requirement
to not do anything special while there could be devices in
suspend-state. Now we are able to see differnce between critical section
with or without suspended devices. For udev synchronization only
suspended devices are prohibited to be there - so slightly relax
condition and allow calling and using 'fs_sync()' even inside critical
section - but there must not be any suspended device.
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.
Internal detection of SCSI device being in-use by DM mpath has been
performed several times for each component device - this could be
eventually racy - so instead when we do remember 1st. checked result
for device being mpath and use it consistenly over the filter runtime.
Move DM usage into dev_manager.c source file.
Also convert STATUS to INFO ioctl - as that's enough
to obtain UUID - this also avoid issuing unwanted flush on checked DM
device for being mpath.
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.
This is the default bcache size that is created at the
start of the command. It needs to be large enough to
hold a single copy of metadata for a given VG, or the
VG cannot be read or written (since the entire VG would
not fit into available memory.)
Increasing the default reduces the chances of anyone
needing to increase the default to use their VG.
The size can be set in lvm.conf global/io_memory_size;
the lower limit is 4 MiB and the upper limit is 128 MiB.
