This message was deleted.

I think this might be the cultprint: https://docs.harvesterhci.io/v1.6/upgrade/v1-4-1-to-v1-4-2#4-virtual-machines-that-use-migratable-rwx-volumes-restart-unexpectedly The two issues: https://github.com/harvester/harvester/issues/8534 - in harvester https://github.com/longhorn/longhorn/issues/11158 - in longhorn for that version we still use longhorn with the problem - 1.4.1. Take a look whether the csi plugin restarts and does that trigger what you see. For the workaround if it does - newer versions of harvester adopted version 1.9.1 of longhorn which I think contains fix for this.

Hi We have observed that the csi-driver pods have restarted multiple times similar to the github issue. As this seems to be a known issue in the longhorn 1.7 version, is it possible to keep the Harvester version as it is and upgrade the longhorn version to a more stable version ≥ v1.8.3 ? Since we are running some critical workloads, it is not possible for us to immediately upgrade the harvester version.

As this is a well known issue for that version and it's fixed in the next harvester version adopting newer longhorn I'd suggest the upgrade is the most reliable approach. The whole stack (longhorn + harvester) is tested against specific versions so upgrading the longhorn version alone to something that works can be unpredictable and is more risky if you truly run critical workloads