Hello
We have the problen that randomly one osd is going into very high latency (100-200ms) while te others stay below 2ms (all ssd micron max 5100) .
After a restart of the osd (another problem this always takes 30 minutes after long time running and only seconds after restarting directly again after a osd restart) it work normal again.
What we also see is that the osd has heavy writes (100 mb/sec+) while it has that high latency. but this can take for a long time. We did not try how many hours this will stay like this because the cluster gets slow because of this osd. Also there is no rebalancing etc at this moment.
Ceph 16.2.7, Kernel Version Linux 5.13.19-4-pve #1 SMP PVE 5.13.19-, PVE 7.1
Osds deployed via Proxmox
Thanks
We have the problen that randomly one osd is going into very high latency (100-200ms) while te others stay below 2ms (all ssd micron max 5100) .
After a restart of the osd (another problem this always takes 30 minutes after long time running and only seconds after restarting directly again after a osd restart) it work normal again.
What we also see is that the osd has heavy writes (100 mb/sec+) while it has that high latency. but this can take for a long time. We did not try how many hours this will stay like this because the cluster gets slow because of this osd. Also there is no rebalancing etc at this moment.
Ceph 16.2.7, Kernel Version Linux 5.13.19-4-pve #1 SMP PVE 5.13.19-, PVE 7.1
Osds deployed via Proxmox
Thanks
Last edited: