Edge case: two backups running at same time on the same Virtual server account causes 2nd backup process to fail

Hi Guys,

I just observed this error today, when there are two backups scheduled at the same time. The first backup creates a .backup directory in the Virtual server account (example.com) and then the second backup process tries to create this .backup directory and aborts with an error.

This would be pretty rare, but in this case the customer wants two backups performed on their website to two different locations and the default schedule of midnight was selected for both backups.

Not sure if this warrants a change, just reporting it.

Thanks ;)

Status: 
Closed (fixed)

Comments

What was the exact error message that you got?

I may have to disallow concurrent backups using the home-style format to prevent this error..

In practice you could work around the problem by specifying two backup destinations for a single backup, which is also more efficient..

Actually, I see a better solution - just wait until one backup is done with the .backup directory for a domain before going on to the next one. I will implement this in the next Virtualmin release.

Thanks Jamie.

I did have a go at scheduling the 2 backup destinations for the same scheduled backup.

However, 1 destination is FTP and the other destination is SCPONLY. And unfortunately, this backup option "Delete old backups" fails for SCPONLY destinations and would cause an backup error to be reported.

And so 2 backup schedules must be created.

It would be handy if the "Delete old backups" option could be set on a per destination basis.

Actually, deletion of old backups should still work for the FTP location, just fail for the SSH destination ..

But in failing for the SCPONLY destination, wouldn't the backup report either a failure or partial success status? Either would be a problem as it is a false positive, since the backup did work.

So do you know what the backup status is set to if the "Delete old backups" action fails?

Oh, you are right .. if any purge of old backups fails, the whole backup will be reported as having failed.

And if you monitoring system is checking the backup status, then it would report this as an error.

So again, it would be handy if the "Delete old backups" option could be set on a per destination basis. :)

I guess that could be done .. but I'd recommend just using two different scheduled backups, once I fix the original bug.

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