Local backups failing

Since updating to 3.70-2 scheduled backups that are stored locally fail. Looking at the backup log I can see that everything is working up until creating the archive with this error: Creating final backup archive .. .. archive failed!

All of our backups are now failing with the same error. Each backup worked up until 3.70-2 was installed July 5 and the backups have not been changed since.

3.70-3 was also installed this morning after reading another support ticket regarding ssh backups. I forced a backup to run and it still fails for the same reason.

Chris

Status: 
Closed (fixed)

Comments

Were these scheduled backups created as root, or as a domain owner?

Also, which backup format did you select (one big file, or one file per domain) ?

The scheduled backups were created as a domain owner.

They are both full and incremental backups. All are single-file with strftime-style time substitutions.

As I mentioned, they worked up until this morning when the schedule backup ran, i.e. after the update. Prior to that, they've run for about two months solid, never with any problem.

There was some changes in this area in 3.70 that may have caused issues.

What was the destination for the scheduled backups?

Also, do scheduled backups created by root work OK?

The destination is local, under virtualmin-backup/ . I won't be able to tell about backups created by root as I do not have that level of access. Maybe by Wednesday I can answer that, though, it'd be great to have this fixed before then as we already have people biting their nails now that their backups are failing.

Actually, Jaimie, looking at the files in the virtualmin-backup directory I noticed something that may be of use. All of the backups before today have the domain administrator ownership. The backup that ran today and failed has root ownership.

Chris

Ok, I see the cause of this - the fix is to replace the file /usr/share/webmin/virtual-server/backups-lib.pl with the file attached to this email, then run /etc/webmin/restart

The fix will also be in the next release.

Automatically closed -- issue fixed for 2 weeks with no activity.