It will be indeed be interesting to see what other people get. My rig has no SSDs, decent quality SATA drives and showed the issue consistently since the availability of the 6.5 kernel. My workload is light, but quite consistent and with a typical workload I had IO loads of 40-60% whereas now (and with 6.2) they are (were) typically 2-3%. So for me, the issue really did show up with a pretty much idle load.
Proxmox VE 8.2.2 - High IO delay
- Thread starter silverstone
- Start date
A shame - mine is still working fine after several days of typical use on both nodes. That's no use to you I know and I guess the problem is complex.
when taking a snapshot the system freezes so much that only a hard restart saves it, and after qm unlock number.
I used a snapshot on an LXC (Home Assistant Core) and it was very quick with no delays. Ditto for a larger W 2102 VM. I appreciate that's not immediately useful for you. Good luck!
Hello Everyone.
Im sorry to say that, using PVE 8.3.1 with kernel 6.8.12-4 , when i restore a a KVM host, all proxmox VMs became unresponsive because of high IO
Im sorry to say that, using PVE 8.3.1 with kernel 6.8.12-4 , when i restore a a KVM host, all proxmox VMs became unresponsive because of high IO