Package: lvm2
Version: 2.02.127-1
Severity: normal
Dear Maintainer,
Context and issue
-----------------
I want to use lvmcache to speedup access to data on RAID1.
Configuration is essentially:
- 2 HDD in software raid1 with md -> one pv -> one VG
- 1 SSD (/dev/sda1) is put in same VG
- 1 of the LV is for /
- I created a lvm cache for /
It happenned /dev/sda1 failed. Boot failed. I booted from Debian Live
CD, but I did not manage to repair. lvtools and cache_check refused to
repair cmeta LV, and I did not find how to access it to try any manual
repairing.
I had to reinstall/restore.
Reproduction
------------
sda1 is the only partition on SSD
sde1 is the only partition on a USB key
*** LV creation
# vgcreate vgtest /dev/sda1 /dev/sde1
# lvcreate -n test -L 10G vgtest /dev/sda1
# lvcreate -n testcache -L 1G vgtest /dev/sde1
# lvcreate -n testcachemeta -L 8M vgtest /dev/sde1
# lvconvert --type cache-pool --poolmetadata vgtest/testcachemeta\
--cachemode writethrough vgtest/testcache
# lvconvert --type cache --cachepool vgtest/testcache vgtest/test
*** LV failure
# lvs -a
# lvchange -an vgtest/test
# pvremove /dev/sde1 --force --force
*** LV recovery tentatives
# lvchange -ay vgtest/test
WARNING: Device for PV NubIyb-joYP-9Jei-Fhy3-xUuz-5SVf-3NSI0I not found or rejected by a filter.
Refusing activation of partial LV vgtest/test. Use '--activationmode partial' to override.
# lvchange -ay vgtest/test --activationmode partial
PARTIAL MODE. Incomplete logical volumes will be processed.
WARNING: Device for PV NubIyb-joYP-9Jei-Fhy3-xUuz-5SVf-3NSI0I not found or rejected by a filter.
Check of pool vgtest/testcache failed (status:1). Manual repair required!
# lvchange -ay vgtest/test_corig --force --activationmode partial
PARTIAL MODE. Incomplete logical volumes will be processed.
WARNING: Device for PV NubIyb-joYP-9Jei-Fhy3-xUuz-5SVf-3NSI0I not found or rejected by a filter.
Unable to change internal LV test_corig directly
# lvconvert -v --force --uncache vgtest/test
WARNING: Device for PV NubIyb-joYP-9Jei-Fhy3-xUuz-5SVf-3NSI0I not found or rejected by a filter.
There are 1 physical volumes missing.
Cannot change VG vgtest while PVs are missing.
Consider vgreduce --removemissing.
There are 1 physical volumes missing.
# vgreduce -v --force --removemissing vgtest
Finding volume group "vgtest"
WARNING: Device for PV NubIyb-joYP-9Jei-Fhy3-xUuz-5SVf-3NSI0I not found or rejected by a filter.
There are 1 physical volumes missing.
There are 1 physical volumes missing.
Trying to open VG vgtest for recovery...
Found same device /dev/sda1 with same pvid SFN1I61m6AOWxnMG5YjWvjVpPuQpNLAc
WARNING: Device for PV NubIyb-joYP-9Jei-Fhy3-xUuz-5SVf-3NSI0I not found or rejected by a filter.
There are 1 physical volumes missing.
There are 1 physical volumes missing.
Archiving volume group "vgtest" metadata (seqno 7).
Removing partial LV test.
Releasing logical volume "testcache"
There are 1 physical volumes missing.
Creating volume group backup "/etc/lvm/backup/vgtest" (seqno 8).
Logical volume "testcache" successfully removed
Releasing logical volume "test"
There are 1 physical volumes missing.
Creating volume group backup "/etc/lvm/backup/vgtest" (seqno 9).
Logical volume "test" successfully removed
Removing PV with UUID NubIyb-joYP-9Jei-Fhy3-xUuz-5SVf-3NSI0I from VG vgtest
Creating volume group backup "/etc/lvm/backup/vgtest" (seqno 10).
Wrote out consistent volume group vgtest
All LV is gone!
Using a cachethrough mode, OriginLV (/dev/sda1) was intact.
There must be some way to access it!
I did not find much clues on Internet.
Best hit: https://www.redhat.com/archives/linux-lvm/2015-August/msg00008.html
Best regards,
Benoit
-- System Information:
Debian Release: stretch/sid
APT prefers testing
APT policy: (990, 'testing'), (500, 'unstable'), (500, 'stable'), (1, 'experimental')
Architecture: amd64 (x86_64)
Foreign Architectures: i386
Kernel: Linux 4.1.0-1-amd64 (SMP w/4 CPU cores)
Locale: LANG=fr_FR.UTF-8, LC_CTYPE=fr_FR.UTF-8 (charmap=UTF-8)
Shell: /bin/sh linked to /bin/dash
Init: systemd (via /run/systemd/system)
Versions of packages lvm2 depends on:
ii dmeventd 2:1.02.104-1
ii dmsetup 2:1.02.104-1
ii init-system-helpers 1.23
ii initscripts 2.88dsf-59.2
ii libc6 2.19-19
ii libdevmapper-event1.02.1 2:1.02.104-1
ii libdevmapper1.02.1 2:1.02.104-1
ii liblvm2app2.2 2.02.127-1
ii libreadline5 5.2+dfsg-3
ii libudev1 225-1
ii lsb-base 4.1+Debian14
lvm2 recommends no packages.
Versions of packages lvm2 suggests:
ii thin-provisioning-tools 0.3.2-1
-- no debconf information
Acknowledgement sent
to Matt Taggart <[email protected]>:
Extra info received and forwarded to list. Copy sent to Debian LVM Team <[email protected]>.
(Sun, 20 Dec 2015 20:33:12 GMT) (full text, mbox, link).
Subject: RE: lvm2: OriginLV lost if CacheLV failed
Date: Sun, 20 Dec 2015 12:28:08 -0800
Hi,
#798431 was filed against version 2.02.127-1, but looking at the upstream
changelog at
https://git.fedorahosted.org/cgit/lvm2.git/tree/WHATS_NEW
I see lots of cache related changes in newer releases. Maybe someone can
try to repeat this bug with the 2.02.133-2 package in stretch/sid or if the
package could be updated to the latest 2.02.138 upstream release (just
released on Dec 14) then you could test with that.
--
Matt Taggart
[email protected]
Debbugs is free software and licensed under the terms of the GNU General
Public License version 2. The current version can be obtained
from https://bugs.debian.org/debbugs-source/.