I just ran pveperf on /mnt/pve/foo & got some horrible numbers, with fsyncs between 17-400.
These disks don't feel anywhere near that bad when running a VM from them.
I imagine I'm simply running the test in a manner outside it's scope.
Is there a standardized proper method by which to quantify & record their performance?
These disks don't feel anywhere near that bad when running a VM from them.
I imagine I'm simply running the test in a manner outside it's scope.
Is there a standardized proper method by which to quantify & record their performance?