Hello !
Can anybody help me with some trouble:
jack at opensolaris:~# zpool status -v
pool: green
state: ONLINE
status: One or more devices has experienced an error resulting in data
corruption. Applications may be affected.
action: Restore the file in question if possible. Otherwise restore the
entire pool from backup.
see: http://www.sun.com/msg/ZFS-8000-8A
scrub: scrub in progress for 0h8m, 5.08% done, 2h42m to go
config:
NAME STATE READ WRITE CKSUM
green ONLINE 0 0 117
c3d0 ONLINE 0 0 234
c4d0 ONLINE 0 0 234
errors: Permanent errors have been detected in the following files:
/green/home/epiq/vid/2resolume/400x300/people/deti/primorsk - manoly -
rebenok na vode.avi
/green/home/epiq/mus/!!![Labels]!!!/playhouse/[playcd007] captain
comatose - going out/11_captain comatose - $100 (live).mp3
green/home:<0x1451a>
green/home:<0x1cd29>
green/home:<0x14537>
green/home:<0x1454e>
green/home:<0x14577>
green/home:<0x1458e>
green/home:<0x14599>
/green/home/epiq/vid/!!!incoming/Mar`ja_Iskusnica.avi
/green/home/epiq/mus/easy
green/home:<0x144ec>
green/home:<0x144f9>
green/home:<0x144fc>
As you can see in output my zfs pool have permanent errors with some files and
directories - but how i can clear these errors? I try to do scrub one time, and
bigger part of errors gone, but some stay, and i can''t delete it :
jack at opensolaris:~# rm -rf /green/home/epiq/mus/easy
rm: Unable to remove directory /green/home/epiq/mus/easy: Directory not empty
jack at opensolaris:~# rmdir --ignore-fail-on-non-empty
/green/home/epiq/mus/easy
rmdir: illegal option -- ignore-fail-on-non-empty
Usage: rmdir [-ps] dirname ...
jack at opensolaris:~# rm -rf
"/green/home/epiq/vid/2resolume/400x300/people/deti/primorsk - manoly -
rebenok na vode.avi"
jack at opensolaris:~# ls -la
"/green/home/epiq/vid/2resolume/400x300/people/deti/primorsk - manoly -
rebenok na vode.avi"
-rw-r--r-- 1 101 staff 18221286 Oct 25 2008
/green/home/epiq/vid/2resolume/400x300/people/deti/primorsk - manoly - rebenok
na vode.avi
While i googled around this problem i found zdb commands:
jack at opensolaris:~# zdb -d green/home 0x1451a 0x1cd29 0x14537 0x1454e 0x14577
0x1458e 0x14599 0x144ec 0x144f9 0x144fc
Dataset green/home [ZPL], ID 40, cr_txg 325, 906G, 116136 objects
Object lvl iblk dblk dsize lsize %full type
83226 2 16K 128K 12.6M 12.6M 100.00 ZFS plain file
118057 4 16K 128K 5.22G 5.22G 100.00 ZFS plain file
83255 2 16K 128K 7.63M 7.62M 100.00 ZFS plain file
83278 2 16K 128K 4.25M 4.25M 100.00 ZFS plain file
83319 2 16K 128K 5.38M 5.38M 100.00 ZFS plain file
83342 2 16K 128K 5.50M 5.50M 100.00 ZFS plain file
83353 2 16K 128K 6.13M 6.12M 100.00 ZFS plain file
83180 2 16K 128K 8.26M 8.25M 100.00 ZFS plain file
83193 2 16K 128K 6.01M 6.00M 100.00 ZFS plain file
83196 2 16K 128K 8.63M 8.62M 100.00 ZFS plain file
, but how it can halp me?
With best wishes, Epiq.
--
This message posted from opensolaris.org
Hi-- The best approach is to correct the issues that are causing these problems in the first place. The fmdump -eV commnand will identify the hardware problems that caused the checksum errors and the corrupted files. You might be able to use some combination of zpool scrub, zpool clear, and removing these corrupted files manually, but unless the failing disks (?) are repaired, you could potentially lose more data. We have zpool recovery feature in build 128 but it will not work in this scenario since the pool is online (its importable). If you can''t repair the underlying hardware issues and you have no backup of this data, then you might consider reviewing the steps that are described here: http://www.solarisinternals.com/wiki/index.php/ZFS_forensics_scrollback_script We haven''t had a chance to look at this script closely or test it yet so the usual caveats apply. Thanks, Cindy On 01/12/10 07:11, epiq wrote:> Hello ! > > Can anybody help me with some trouble: > > jack at opensolaris:~# zpool status -v > pool: green > state: ONLINE > status: One or more devices has experienced an error resulting in data > corruption. Applications may be affected. > action: Restore the file in question if possible. Otherwise restore the > entire pool from backup. > see: http://www.sun.com/msg/ZFS-8000-8A > scrub: scrub in progress for 0h8m, 5.08% done, 2h42m to go > config: > > NAME STATE READ WRITE CKSUM > green ONLINE 0 0 117 > c3d0 ONLINE 0 0 234 > c4d0 ONLINE 0 0 234 > > errors: Permanent errors have been detected in the following files: > > /green/home/epiq/vid/2resolume/400x300/people/deti/primorsk - manoly - rebenok na vode.avi > /green/home/epiq/mus/!!![Labels]!!!/playhouse/[playcd007] captain comatose - going out/11_captain comatose - $100 (live).mp3 > green/home:<0x1451a> > green/home:<0x1cd29> > green/home:<0x14537> > green/home:<0x1454e> > green/home:<0x14577> > green/home:<0x1458e> > green/home:<0x14599> > /green/home/epiq/vid/!!!incoming/Mar`ja_Iskusnica.avi > /green/home/epiq/mus/easy > green/home:<0x144ec> > green/home:<0x144f9> > green/home:<0x144fc> > > As you can see in output my zfs pool have permanent errors with some files and directories - but how i can clear these errors? I try to do scrub one time, and bigger part of errors gone, but some stay, and i can''t delete it : > > > jack at opensolaris:~# rm -rf /green/home/epiq/mus/easy > rm: Unable to remove directory /green/home/epiq/mus/easy: Directory not empty > jack at opensolaris:~# rmdir --ignore-fail-on-non-empty /green/home/epiq/mus/easy > rmdir: illegal option -- ignore-fail-on-non-empty > Usage: rmdir [-ps] dirname ... > jack at opensolaris:~# rm -rf "/green/home/epiq/vid/2resolume/400x300/people/deti/primorsk - manoly - rebenok na vode.avi" > jack at opensolaris:~# ls -la "/green/home/epiq/vid/2resolume/400x300/people/deti/primorsk - manoly - rebenok na vode.avi" > -rw-r--r-- 1 101 staff 18221286 Oct 25 2008 /green/home/epiq/vid/2resolume/400x300/people/deti/primorsk - manoly - rebenok na vode.avi > > While i googled around this problem i found zdb commands: > > jack at opensolaris:~# zdb -d green/home 0x1451a 0x1cd29 0x14537 0x1454e 0x14577 0x1458e 0x14599 0x144ec 0x144f9 0x144fc > Dataset green/home [ZPL], ID 40, cr_txg 325, 906G, 116136 objects > > Object lvl iblk dblk dsize lsize %full type > 83226 2 16K 128K 12.6M 12.6M 100.00 ZFS plain file > 118057 4 16K 128K 5.22G 5.22G 100.00 ZFS plain file > 83255 2 16K 128K 7.63M 7.62M 100.00 ZFS plain file > 83278 2 16K 128K 4.25M 4.25M 100.00 ZFS plain file > 83319 2 16K 128K 5.38M 5.38M 100.00 ZFS plain file > 83342 2 16K 128K 5.50M 5.50M 100.00 ZFS plain file > 83353 2 16K 128K 6.13M 6.12M 100.00 ZFS plain file > 83180 2 16K 128K 8.26M 8.25M 100.00 ZFS plain file > 83193 2 16K 128K 6.01M 6.00M 100.00 ZFS plain file > 83196 2 16K 128K 8.63M 8.62M 100.00 ZFS plain file > > > , but how it can halp me? > > With best wishes, Epiq.
Cindys, thank you for answer, but i need explain some details. This pool is new hardware for my system - 2x1Tb WD Green hard drives, but data on this pool was copied from old 9x300 Gb hard drives pool with hw problem. while i copied it data where was many errors, but at the end i see this picture : jack at opensolaris:~# fmdump -eV fmdump: failed to open /var/fm/fmd/errlog: No such file or directory and where no READ or WRITE errors on new pool - only CKSUM. As i understand it closely to transfer errors, then problem with new HW. And now i need only to clear this permanent errors on new pool, for probably restoring this files from backups. -- This message posted from opensolaris.org
Hi, I think you are saying that you copied the data on this system from a previous system with hardware problems. It looks like the data that was copied was corrupt, which is causing the permanent errors on the new system (?) The manual removal of the corrupt files, zpool scrub and zpool clear might work, but I don''t have experience with this many errors on a non-redundant config. If you have a clean backup of this data, you might consider destroying the green/home dataset, clearing the pool errors, recreating green/home, and restoring the known, good data. Maybe someone else can suggest a better approach... Cindy On 01/12/10 15:27, epiq wrote:> Cindys, thank you for answer, but i need explain some details. This pool is new hardware for my system - 2x1Tb WD Green hard drives, but data on this pool was copied from old 9x300 Gb hard drives pool with hw problem. while i copied it data where was many errors, but at the end i see this picture : > jack at opensolaris:~# fmdump -eV > fmdump: failed to open /var/fm/fmd/errlog: No such file or directory > > and where no READ or WRITE errors on new pool - only CKSUM. As i understand it closely to transfer errors, then problem with new HW. And now i need only to clear this permanent errors on new pool, for probably restoring this files from backups.