Illegal byte sequence after disk replace

[fede2222]

Hi

I get this message after a disk replace... i hope you help me.

I got my zpool degraded... the zpool show one of the disks unavailables, i get it out, run SMART and surfaces checks and the disk is allright.
So i reinitialize the disk (clear firsts and lasts sectors). I put the disk, and make a zpool replace. The resilver was finished but with these:

pool: media2
state: DEGRADED
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: resilver completed after 0h26m with 277 errors on Wed Feb 20 18:43:22 2013
config:

NAME STATE READ WRITE CKSUM
media2 DEGRADED 0 0 978
raidz1 DEGRADED 0 0 3.72K
gptid/93b44ead-e4f3-11e1-9ac1-bcaec528164a ONLINE 5 685 0
gptid/941a2da1-e4f3-11e1-9ac1-bcaec528164a ONLINE 0 0 0
replacing DEGRADED 0 0 0
14408424784145491985 UNAVAIL 0 0 0 was /dev/gptid/a69803c7-fba1-11e1-a7a6-bcaec528164a
ada6p2 ONLINE 0 0 0 112G resilvered
gptid/94eb7026-e4f3-11e1-9ac1-bcaec528164a ONLINE 0 0 0

errors: 277 data errors, use '-v' for a list


With -v it says some files was corrupted... but the worst case is that some dirs are missing. When i do ls it says:
ls: Anime: Illegal byte sequence
ls: Anime y Cartoons.xls: Input/output error
ls: Avatar - The Legend Of Korra (2012-): Input/output error
ls: Cartoons: Input/output error
ls: Green Lantern - The Animated Series (2012-): Input/output error
ls: Metalocalypse (2006-): Input/output error
ls: Thundercats (2011-): Input/output error
ls: ____downloads2: Input/output error
./ 3d/ _ Openings _/ _! Videos !_/
../ _ Bajando _/ _ Para Probar _/


Is there a way to recover the files??

Thanks

 

[cyberjock]

Nope. Recover from backup. Now you see why RAIDZ2 is better than RAIDZ1. If you had a RAIDZ2 you'd likely have not lost any data.

The problem with RAIDZ is if 1 disk fails and any other disks have any bad sectors or are otherwise out of sync with the zpool(aka you don't do regular scrubs) then you can expect data loss. Period. Disks are so big these days and relibility is not improving as fast as disk size is increasing so you can statistically almost be guaranteed that you will have at least 1 bad sector on at least 1 disk if your array is bigger than 20TB. Less than 20TB and you still have a really good chance, but it's not quite guaranteed.

 

[fede2222]

The problem is a disk is going out (lost device in dmesg) frequently... but when i run zpool status this disk in the pool is showed online!!!... this is normal???. Because this disk was out, i cannot access the archives -> "illegal byte sequence", that make sense.

After a change of cables and change of sata ports where the HDDs connected, all disks now perfom well. No more randoms "lost device"... It fixed the "illegal byte sequence" and can access the files, one problem solved. Only the errors of ZFS of zpool status persists. Then i made a scrub, it finished, and zpool status don't show errors now. It means that all the files were corrected??? :). I check the files and apparently look ok.

pool: media2
state: ONLINE
scrub: none requested
config:

NAME STATE READ WRITE CKSUM
media2 ONLINE 0 0 0
raidz1 ONLINE 0 0 0
gptid/93b44ead-e4f3-11e1-9ac1-bcaec528164a ONLINE 0 0 0
gptid/941a2da1-e4f3-11e1-9ac1-bcaec528164a ONLINE 0 0 0
ada6p2 ONLINE 0 0 0
gptid/94eb7026-e4f3-11e1-9ac1-bcaec528164a ONLINE 0 0 0

errors: No known data errors


I know is better to go to RAIDZ2, i bought the disks and going to that on the weekend.

 

[cyberjock]

It is normal for you to see all disks even if they aren't connected. The disks show as online for a little while after a disk is lost because FreeBSD 8 doesn't use zfsd to identify a missing disk the instant it is disconnected.

If you did a scrub and it says "No known data errors" then any problems(if there were any) it found were repairable.

 

[fede2222]

ok... thanks for the reply and the tips... :)