FYI, I had to edit /usr/bin/vzdump to change $lvcreate --size 500m to $lvcreate --size 2000m. My vzdump was failing multiple days in a row on my Plesk jail. It has about 60 domains in it and about 30G used, nothing really out of the ordinary. 500m just wasn't enough and it failed and started spewing filesystem errors after about 50 minutes. It's happy now at 2000m.
vzdump lvm snapshot size
- Thread starter tog
- Start date
A lvm snapshot only records changes. Are you really sure that your VM writes more than 500MB whithin 50 minutes?
- Dietmar
- Dietmar
I don't quite know how 500m worth of changes could be accumulating in under an hour, but apparently that was the case, yes.
For at least 3 or 4 nights in a row after ~50 minutes the vzdump would fail out with nasty critical filesystem errors every single time. I upped the snapshot size from 500m to 2000m and it has been perfectly fine for at least three nightly dumps since.
For at least 3 or 4 nights in a row after ~50 minutes the vzdump would fail out with nasty critical filesystem errors every single time. I upped the snapshot size from 500m to 2000m and it has been perfectly fine for at least three nightly dumps since.