Is there a way to re-nice the backup/restore process? Any time I try to restore a large (+10GB) virtual server backup on EBS-backed EC2 (which suffers from bad i/o performance), the local mysql and apache servers hang and sites start to time out. I’d rather not have to kick off my restores from the command line, so is there any functionality to ‘nice -n 19’ site restores?
You can configure that by going into Webmin -> Webmin -> Webmin Configuration -> Advanced Options, and there, you can set “CPU priority for scheduled jobs”, “IO class for scheduled jobs”, and “IO priority for scheduled jobs”.
Brilliant! It’s incredible how much configuration is possible here. I wish there were some way to organize it all a little more intuitively, but I’d be hard-pressed to reorganize it in a moree sensible way.
That should apply to non-scheduled jobs as well, correct? That means the ‘tar’ application that’s forked should be run with the same i/o priority, correct?
The backup process, and everything spawned by it, should run with the CPU and I/O priority you specify in the settings there.