Stale extent won't go away!

[AIMC]

Hi All,

I had 1 stale extent (the first extent of /dev/vg00/lvol8) but only on one of my mirrored disks, so i replaced the disk, vgcfgrestore etc... got an i/o error saying

"couldn't resynchronise logical volume /dev/vg00/lvol8"

so i still have 1 stale extent on this mirrored copy, i then tried lvreducing all lvols, vgreduce & then pvcreate the disk again & mirror from scratch.... same problem.

Now i think the first extent of /dev/vg00/lvol8 which is showing current on my supposed good disk is actually corrupt, so apart from restoring from ignite is there any other way to make both mirrored extents current & remove the corruption?

Thanks in advance!

Aidan

[Steven E. Protter]

Shalom,

seems the mirror copy has a bad spot on the disk. I'd test and replace the hardware and then restore from Ignite.

SEP

[AIMC]

Is there any way of finding out what data is actually in this extent? If so maybe we could copy the data back in from somewhere else to clear the corruption because i'm pretty sure the disk is good..

Aidan

[AIMC]

Another question...

Could i restore /dev/vg00/lvol8 (/var) over the existing /var filesystem on a running server?

[Peter Nikitka]

Hi,

of course you could do that - but you probably be sorry of having it done!

There are many things in /var, which reflect the state of the OS - but that state is changing permanently (patches, product installations, open files of daemon processes, ...).

Unless the backup is really really close to the curtent system state, I would not consider this as a solution of a (potential) disk problem.

And if you do it (backup really up2date), then only in single user mode.

mfG Peter