Restore from backup

Hey There,

I’ve scheduled a weekly backup from Virtualmin and now i need to restore some files. I’ve made the approrpiate selection but after i click on the “Show what will be restored” button the browser works a little then shows again the restore page. Though the machine is busy with the extraction, but the Virtualmin interface doesn’t show the list of files.
Is this a timeout issue?

did you check any log files?
When I do a restore it goes through without hesitation.

The problem was with my browser. With Firefox it’s working.

Another question: is there a way to restore just single files not whole servers home directory?


While I don’t believe you can restore a single file from within Virtualmin, it’s not too hard to do that from the command line.

If you have a .tar.gz file that looks like this:


And you want the users file named "public_html/index.html" – you could retrieve just that file by typing this:

tar zxvf my.domain.tar.gz public_html/index.html

If you aren’t sure which file you want, you can browse the list of files (without actually extracting any) by typing this:

tar tzvf my.domain.tar.gz

You can restore individual features, but not individual files (except via the method Eric covered). We’ll probably add a command line switch for restoring a single file in the future, though it takes some additional thinking to figure out what you’re asking for–Virtualmin knows things by features and where it puts them on restore is based on where on the restoring system those features go (sounds strange, but it’s a Good Thing), so it would have to calculate what file you mean when you ask for an absolute path to be restored, since it knows it as a path relative to where configuration puts that particular feature (and optionally within a sub-server home rather than a server home). Tricky stuff. But, the format is pretty easy to work with to find things you need, as we try to keep everything accessible, configurable and recoverable via many paths, not just the ones we’ve thought of yet.

The problem was with my browser. With Firefox it's working.

Unfortunately IE has a very short timeout (something like 20 or 30 seconds), which causes problems for long processes. I believe Jamie just checked in some code to add a ticker to this page to keep the connection alive even on browsers that have a short timeout, but I might be misremembering where that change occurred…it was either here or in one of the system backup tools.