This message was deleted.
# longhorn-storagea
adamant-kite-43734
05/01/2024, 5:36 AMThis message was deleted.
c
cool-architect-86201
05/01/2024, 5:39 AMI checked the resource status of pods under the
longhorn-systemq
quick-river-12881
05/01/2024, 3:53 PMThe remount is a feature: https://github.com/longhorn/longhorn/issues/6386
c
cool-architect-86201
05/01/2024, 4:01 PM@quick-river-12881 Oh, I think I understand now. This might be due to network or disk latency.
I want to know how long Longhorn can tolerate latency at most, as we really don't want to see container restarts.
Also, I noticed there are many restarts of the attacher container, and checking the logs, I found 'lost connection.' What could be causing this?
cool-architect-86201
05/01/2024, 4:01 PM2024-05-01T142440.893215718Z E0501 142440.893165 1 connection.go:142] Lost connection to unix:///csi/csi.sock.
2024-05-01T142440.893249648Z F0501 142440.893237 1 connection.go:97] Lost connection to CSI driver, exiting
q
quick-river-12881
05/01/2024, 4:05 PMIs it possibly an instance of https://github.com/longhorn/longhorn/issues/8427 ?
👍 1
p
powerful-librarian-10572
05/01/2024, 6:28 PMI see your data locality is set as disabled. Maybe set your volumes to best effort if you can? If network capacbilties are too constraining for your cluster, this could be a workaround, i believe.
c
cool-architect-86201
05/10/2024, 3:29 AMOh, thank you for the reminder. I was using 'best-effort' in the StorageClass, but it was disabled in the UI panel. I'm not sure which one it's based on. I've now corrected both to 'best-effort'. @powerful-librarian-10572
44 Views