Kernel Panic by SCSI

March 11, 2021 22:43

I’ve notices some kernel panics on one of my vms when ever any other vm was doing some more cpu intensive work. It seemed there was no reason, as there was no unusual ram usage or vm-host events! So I ignored it for a while until I was finally able to trigger it by using one of my own services - here is a little record of the crash (just in case of further reference):

For all which do not want to capture the exact frame where the panic begins - here is a type-off it:

[534.726288] 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

Soo, what is my solution?! It seems this is a bug inside the virtual SCSI controller, which I use for all my disks - as SCSI is the only protocol supporting discarding blocks to the underlying ZFS filesystem I use for the vm disks (further reference is here). The solution was just to switch to the VirtIO SCSI model for the SCSI controller. That’s it. No seemingly random crashes anymore! I hope this helped you.