I think the problem lies in the combination detect_zeroes and discard.
I changed one of my vm:s config-file in /etc/pve... to “detect_zeroes=off,discard=on” and I haven't had any problem with that vm for several days.
Can anybody else confirm this? I don't have so high load on my cluster now so I don't know if it's just a coincidence.
I changed one of my vm:s config-file in /etc/pve... to “detect_zeroes=off,discard=on” and I haven't had any problem with that vm for several days.
Can anybody else confirm this? I don't have so high load on my cluster now so I don't know if it's just a coincidence.