LVM failure while pvmove (help needed urgently)

[Mel Burslan]

During the pvmove from one disk to the other, process quit about 85% of the way
when I restarted pvmove, I started getting IO errors and pvmove got aborted after every extent. I kept repeating (now it looks like a very BAD idea) the pvmove command untill all extents seemed to move. But right now my filesystem is disable. In bdf listing, I see it as

bdf: /app: I/O error

of course it is unmountable and can not be fsck'ed

what are my options ? I am hoping someone here can tell me faster than HP, with whom I am on the phone for the past 25 minutes and I could not get passed to the HPUX team.

[Mel Burslan]

forgot to mention, fsck errors out as follows :

vxfs fsck: file system had I/O error(s) on user data.
vxfs fsck: OLT extent 0 has bad magic
read of primary OLT failed
vxfs fsck: OLT extent 1 has bad magic
read of OLT copy failed
vxfs fsck: no valid OLT, cannot continue
file system check failure, aborting ...

[Ashwani Kashyap]

Can you at least do an lvdisplay on the LV which is mounted on the failed file system .

Is vgidsplay ok . How about the source and the destination disks . were there any bad blocks/sectors on any before you started pvmove .

[Mel Burslan]

lvdisplay output looked kind of crappy. the temporary mirror copies created on the lv was not removed and there was amysterious extent on one of the alternate path volumes...

was a big mess. ate up my whole day. we had to bring the system down to clear it up. now after lvremove, lvcreate, newfs, and a 2 hrs long restore, all is well on the western front.

Thanks.

[Michael Tully]

I've seen this happen (the exact error message) when a LUN has been accessed from another system. What it is, is the actual LVM header (the volume group information etc) and FS have been flattened.