Changing SCSi controller

[manfer]

Hi,

I have Proxmox 7.1-7 with a ZFS volume and a VM with debian 10 OS on a SCSi disk with LSI 53C895A controller. From time to time the vm gets stuck with a kernel panic that mentions LSI driver. Mostly when it is doing the weekly bakcup.

Is it safe to stop the VM, change the controller to Virtio SCSi Single and turn on the VM again? Is there anything to do to make it work after the change or will it just boot properly?

Should I do some other change to improve performance? Like enabling SSD Emulation on the Hard Disc or Discard? And can those options be enabled with no harm to the VM and its data?

Thanks.

 

[gfngfn256]

Is it safe to stop the VM, change the controller

Make sure you have a full & restorable backup of the VM & it's data (test it out before by restoring to a different VMID/system) and there is no risk.

As a rule-of-thumb; you should always have full backups of any VMs/LXCs you use & most definitely before making any changes in their configs/setup.

 

[leesteken]

You are running a Proxmox version that has had many fixes and updates that you chose not to install. Even the latest 7.4 will be out of support at the end of the month.
Who knows how long your VM has been running without a stop and restart. Maybe you changed various configuration settings (over the years) that only take affect when you restart the VM.
Not updating and not restarting for long enough will create their own risks and problems...
In general Linux VM's support VirtIO SCSI out of the box but you might want to test this first with a copy of your VM.

 

[manfer]

Then I must had some repository not enabled as the server does not prompt me to update proxmox.

As for the virtual machine had been stopped and restarted quite recently. The kernel panic forces to do that. Last one was as recently as this weekend.

What the kernel panic says exactly is,
Kernel panic - not syncing: assertion "i && sym_get_cam_status(cp-cmd) == DID_SOFT_ERROR" failed: file "drivers/scsi/sym53c8xx_2/sym_hipd.c", line 3399

Does why I'm considering changing the controller setting.

As for changing any settings without a restart, nothing like that happened.

The ZFS volume is being replicated to a second server in a cluster of those two servers, other that the weekly backup of the virtual machine.

 

[gfngfn256]

As for the virtual machine had been stopped and restarted quite recently

So why the question:

Is it safe to stop the VM

 

[leesteken]

Then I must had some repository not enabled as the server does not prompt me to update proxmox.

Seems I misjudged the question about stopping as an aversion to restart anything, sorry. Maybe just upgrade your Proxmox to the latest 7.4 and see if that fixes the issue?

As for the virtual machine had been stopped and restarted quite recently. The kernel panic forces to do that. Last one was as recently as this weekend.

What the kernel panic says exactly is,
Kernel panic - not syncing: assertion "i && sym_get_cam_status(cp-cmd) == DID_SOFT_ERROR" failed: file "drivers/scsi/sym53c8xx_2/sym_hipd.c", line 3399

Does why I'm considering changing the controller setting.

As for changing any settings without a restart, nothing like that happened.

As I said: In general Linux VM's support VirtIO SCSI out of the box but you might want to test this first with a copy of your VM.