I have recently transitioned my server from Hyper-V to Proxmox 8.3 and encountered an issue with my Windows VM running MSSQL. Here are the server and configuration details:
The issue arises when I use the VirtIO SCSI single controller on the MSSQL VM:
The VM hangs within 24 hours or less, requiring a reboot.
All other Windows VMs continue to operate without issues.
To troubleshoot, I switched the MSSQL VM to the VMware PVSCSI controller, which prevents it from hanging but produces similar warnings as the other Windows VMs. Specifically:
Warnings: Occur nightly between 10–11 PM.
Logs: Include repeated Reset to device warnings (vioscsi for VirtIO or pvscsi for PVSCSI).
Additional context:
The ZFS ARC is set to 64GB.
No backups or scheduled tasks run during the time the warnings appear.
I'm looking for advice on diagnosing or resolving the cause of the hangs and warnings. Could this be related to ZFS, ARC size, or the SCSI controller drivers? Are there known compatibility issues with VirtIO or PVSCSI in this scenario?
The issue you're describing sounds like a known problem with Virtio drivers. It's been identified, debugged, and actively worked on for some time. As far as I know, a fix has been confirmed, although I’m not entirely sure of its current publishing status.
In the meantime, the usual workaround has been to downgrade the drivers to a specific version. However, you might be able to use a version that includes the fix.
