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.
When we tested lvm2, the kernel injected various random faults.
(gdb) bt
...
(gdb) p vg
$1 = (struct volume_group *) 0x0
(gdb) p use_previous_vg
$2 = (unsigned int *) 0x0
Signed-off-by: Wu Guanghao <wuguanghao3@huawei.com>
When compiled and used with:
CFLAGS="-fsanitize=address -g -O0"
ASAN_OPTIONS=strict_string_checks=1:detect_stack_use_after_return=1:check_initialization_order=1:strict_init_order=1
we have few reported issue - they where not normally spotted, since
we were still accessing our own memory - but ouf of buffer-range.
TODO: there is still something to enhance with handling of #orphan vgids
Corrupt metadata text (with good mda header) was being handled
in the label_scan phase, but not in the vg_read phase. This
was sufficient because metadata areas would always be read and
checksummed during label_scan (metadata parsing was skipped
previously as an optimization.)
This changed with the optimization in
commit 61a6f9905e
"metadata: optimize reading metadata copies in scan"
Now, some metadata areas will not be read and checksummed
at all during the label_scan phase, only during the vg_read
phase. This means that bad metadata text may first be detected
in the vg_read phase. So, add equivalent bad metadata handling
to the vg_read path to match the label_scan path.
While being in lockless scanning phase, we can avoid reading and checking
matching metadata copies if we already know them from other PV
and just rely on matching metadata header information.
These copies will be examined later during locked metadata read/write
access.
This patch may postpone discovering some read failures to locked phase.
Preserve computed crc32 check from first written PV, just like we
preserve generated metadata.
Also there is no need to call crc32 twice on wrapping buffer with 2 calcs,
result must be always the same as with single crc32 checking.
pvid and vgid are sometimes a null-terminated string, and
other times a 'struct id', and the two types were often
cast between each other. When a struct id was cast to a char
pointer, the resulting string would not necessarily be null
terminated. Casting a null-terminated string id to a
struct id is fine, but is still avoided when possible.
A struct id is: int8_t uuid[ID_LEN]
A string id is: char pvid[ID_LEN + 1]
A convention is introduced to help distinguish them:
- variables and struct fields named "pvid" or "vgid"
should be null-terminated strings.
- variables and struct fields named "pv_id" or "vg_id"
should be struct id's.
- examples:
char pvid[ID_LEN + 1];
char vgid[ID_LEN + 1];
struct id pv_id;
struct id vg_id;
Function names also attempt to follow this convention.
Avoid casting between the two types as much as possible,
with limited exceptions when known to be safe and clearly
commented.
Avoid using variations of strcpy and strcmp, and instead
use memcpy/memcmp with ID_LEN (with similar limited
exceptions possible.)
There have been two separate checks for metadata
validity: first that the metadata text begins with
a valid VG name, and second the checksum of the
metadata text. These happen in different places,
which means there have been two separate error paths
for invalid metadata. This also causes large metadata
to be read in multiple parts, the first part is read
just to check the vgname, and then remaining parts are
read later when the full metadata is needed.
This patch moves the vg name verification so it's
done just before the checksum verification, which
results in a single error path for invalid metadata,
and causes the entire metadata to be read together
rather that in parts from different parts of the code.
If label_scan encounters bad vg metadata, invalidate
bcache data for the device and reread the mda_header
and metadata text back to back. With concurrent commands
modifying large metadata, it's possible that the entire
metadata area can be rewritten in the time between a
command reading the mda_header and reading the metadata
text that the header points to. Since the label_scan
is just assembling an initial overview of devices, it
doesn't use locking to serialize with other commands
that may be modifying the vg metadata at the same time.
Use different 'hint' size for dm_hash_create() call - so
when debug info about hash is printed we can recognize which
hash was in use.
This patch doesn't change actual used size since that is always
rounded to be power of 2 and >=16 - so as such is only a
help to developer.
We could eventually use 'name' arg, but since this would have changed
API and this patchset will be routed to libdm & stable - we will
just use this small trick.
Every vg_write stores new 'metadata' into precommitted slot.
For this step we use 'serialized buffer' to ascii metadata.
Instead of recreating this buffer after whole 'vg_write()' we
use this buffer instantly for creating of precommitted VG.
This has also the advantage of catching any problems with
reparsing of ascii metadata back to VG early before any write.
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.)
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>
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.)
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.
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 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.)
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
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.
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.
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.
When a single copy of metadata gets within 1MB of the
current io_memory_size value, begin printing a warning
that the io_memory_size should be increased.
. 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.
lvm uses a bcache block size of 128K. A bcache block
at the end of the metadata area will overlap the PEs
from which LVs are allocated. How much depends on
alignments. When lvm reads and writes one of these
bcache blocks to update VG metadata, it can also be
reading and writing PEs that belong to an LV.
If these overlapping PEs are being written to by the
LV user (e.g. filesystem) at the same time that lvm
is modifying VG metadata in the overlapping bcache
block, then the user's updates to the PEs can be lost.
This patch is a quick hack to prevent lvm from writing
past the end of the metadata area.
Previously the size was limited by checking if the
old and new copies of the metadata overlapped.
This generally limited the size to about half of
the total space, but it could be larger given the
size differences between old and new. Now add a
direct check to limit the size to half the space.
