I am trying to do an automated Virtualmin backup, but consistently receive an error.
The specifics on the backup are:
Incremental, All Virtual Servers - Include sub-servers, all features, backup to a local file (the directory exists), single archive file, email on failure, run on simple hourly schedule.
The job has been scheduled, and runs on schedule, however after moving through all servers with no errors, it finishes with the following:
Creating final backup archive ..
.. archive failed!
Backup failed! See the progress output above for the reason why. Total backup time was 00 minutes, 50 seconds.
The “progress output above” shows no errors with the exception of the “…archive failed!” which really doesn’t give me much information. Any one know what “archive failed” means? If I run the same backup job in real-time, it completes successfully.
Thanks for all help!
Which version of Virtualmin are you using? One or two versions ago, there was a bug where it wasn’t producing any debug output (such as the above).
Even though it says “failed”, does it still create an archive file?
Also, if you change the backup type from “Single archive file” to “New Format”, which would use one file per Virtual Server, does that make a difference?
My Virtualmin version is 3.72.gpl.
The scheduled backup creates an archive file in the proper directory, named correctly according to the scheduled job, but with a size of 0 bytes.
I have changed the type to “New Format” and should see the results in about 15 minutes, and will update at that time.
When I set the backup as a new format, individual files, it works. This is not ideal, however since I have several servers/domains, and I am primarily creating these backups in case of hardware failure, when I’d need to restore all the servers at once to a backup machine.
Any other thoughts?
Well, with individual files, you have the option of restoring one at a time, or all at once.
To restore all at once, you’d simply point Virtualmin to the directory they reside in, and Virtualmin could restore that as quickly as it could one large file.
In that regard, there really isn’t an advantage of using one large file.
That said – one large file should indeed work, if your /tmp directory isn’t restrained by available disk space.
If disk space isn’t a problem, and you really want one large file, you may want to file a bug report using the Support link above.
Thanks Eric, I didn’t realize Virtualmin could restore all servers at once…or not when they are backed up individually. That really does make the “individual files” option significantly more flexible and useful than one large file in my mind.
I’ll just leave this a sleeping dog and let it lie then.
I’ve noticed that backups are much more successful for virtual servers which pass the verification tests