File system corruption

[John Kelly_3]

I'm trying to recover from a filesystem corruption situation. We were getting inode errors in the syslog from a filesystem and by excluding the faulty files, I managed to run an fbackup. I had the disk replaced and recreated the filesystem. However when I frecover, the filesystem is corrupt.
fsck -y -o full -y /dev/vg01/lvol2
doesn't fix it.
I get loads of
fileset 999 primary inode 95306 unreferenced file, reconnect? (ynq)y
fileset 999 primary inode 95307 unreferenced file, reconnect? (ynq)y
fileset 999 primary inode 95308 unreferenced file, reconnect? (ynq)y

Any suggestions?
It is on HPUX 11.0, A500-7X.

[Steven E. Protter]

Shalom,

newfs the filesystem
Restore all data.

SEP

[John Kelly_3]

I'm getting these errors AFTER I've
replaced the disk,
created the LV
run newfs
frecover'ed the data

Any suggestions?

[siva0123]

Johan,

Why cant you try an fsck with an alternate superblock copy.

Alternate superblock copies will be availble
in /etc/sbtab

then do a fsck with -b option i hope .

Thanks,
Siva

[siva0123]

The command should be something like this.

fsck -F filesystemtype -b alternatesuperblockcopy -o full


Thanks,
Siva

[John Kelly_3]

I'll give the fsck suggestion a try once I finish another restore.
First restore was onto a mirrored LV. - faulty
I split the mirror...
Second restore was to LV on new disk. - faulty
Trying onto the second disk now.
During previous restores there were errors like this (which was why I ran fsck after the restore) -
vxfs: mesg 008: vx_direrr - /volumes file system inode 16195 block 1946498 error 6

Still no disk errors and there weren't any stale extents.

[siva0123]

John,

While you do the restore ,

Please go thorugh this thread.

http://forums1.itrc.hp.com/service/forums/questionanswer.do?threadId=945777

Thanks,
siva

[John Kelly_3]

fsck -F vxfs -b alternatesuperblockcopy -o full /dev/vg01/lvol2
fsck: illegal option -- b


:-(

Still running the third restore

[siva0123]

John,

I hope you used the following command

fsck -F vxfs -b XXX -o full /dev/vg01/rlo\vol2

where XXX=>Alternate superblock copy no which you got from /etc/sbtab file

If that have failed i'm having a feeling that -b option is supported for only root filesystem /dev/vg00/rlvol1 .

I've tested this before successfully in /dev/vg00/lvol1

Let me confirm.

Thanks,
siva

[John Kelly_3]

Ok, I finished the restore onto the last remaining disk and it was successful with no filesystem errors. So the unlikely conclusion is that the replacement disk was as faulty as the original one.
That is thanks to the cheap third-party hardware support company.

Genuine thanks to all who responded for your time.

[John Kelly_3]

Replacement disk was faulty too. Restored onto another disk.