Scalix Forums

hi!

system-description:
suse enterprise server 9 sp3. scalix residents in a lvm (ext3) with mountpoint /var/opt/scalix.
we do a offline backup every night (omshut, backup, omrc). the "backup-software" take a copy of the /var/opt/scalix directory.

problem:
since friday it is not possible to copy some directories under /var/opt/scalix/data. we got read/write errors!!

question
which steps we have to do to solve this problem?
could it be a problem of the lvm or ext3?
is it possible to solve the problem with scalix commandline tools?

thanx 4 helping me!

Sounds to me like corruption on the disk, due to drive failure or filesystem failure (could be either).

How to recover from this would depend on what you need to do. If you can live with older data (most can't) you can just restore an older backup.

If you want to recover corrupted data you will have to first make sure the drive is fine (copy the data to a new disk), then run a omscan -A -a -f -x to correct the mailstore (this will probably delete some mails/folders/mailboxes). You can then recover the deleted mails from one of your older (correct) backups and import the missing data back into the mailstore.

Not the easiest thing in the world. For this I would recommend you get in touch with Scalix and ask them to help you recover this.

thank u valerion,

i think it is a corruption of the ext3 filesystem. the harddisks (raid 5) are definitly ok. i will do a fsck and hope this will solve my problem.

then i will try to fix the message store. sound this good for you?

Sounds fair to me, yeah. But a word of warning, mailstore recovery is possible, and using the tools correctly does allow you to do many things, but its not always easy, and will take up quite some time. You will most probably lose very little, only the mails / attachments that are actually on the affected portions of the filesystem. Those will probably be lost. But, since you have backups, it will help at least some.

But I suggest you do a deeper investigation here. ext3fs does not, in my experience, just corrupt on its own. Other factors play a big role (unclean shutdowns, corruption by the RAID controller, physical errors, etc). If this was my site, after recovery I would try to find out why the corruption started, to prevent any future re-occurances.