if using the virtio-blk drivers version 1.189-1 the VM exits with:
bhyve exited with status 134
If using the virtio-blk drivers version 0.1.141 are used nothing crashes
but the disk is not found...
Any ideas or how to debug this further?
The "stable" version (0.1.185) from fedora works for me, you can find it linked on the following page:
Look for "Stable virtio-win iso".
many thanks for the hint.
Yes, this version is working fine (for some time).
Unfortunately I get a bluescreen in Windows and:
pid 39852 (bhyve), jid 0, uid 0: exited on signal 6
when running a CrystalDiskMark benchmark on this disk :-(
So at least its not finally stable somewhere.
Could you try to stress the virtio-blk disks a bit.
BTW if its relevant:
I put the virtio-blk disk on a ZFS vol with 4k blocksize.
There is a known issues with 1.189. https://bugzilla.redhat.com/show_bug.cgi?id=1890810
If you are either presenting the zvol as 4k or the underlying setting of it is 4k, then you'll run into corruption pretty quickly.
I have tested and confirmed this. 1.187 does not exhibit the same issue. Testing was conducted using 12.2-RELEASE and Windows 10.
Our Windows Server 2019 fleet are using 1.185 without issue.
(In reply to Jason Tubnor from comment #3)
What blocksize would you recommend for NTFS?
The ZFS pool is on a 4k SSD.
I tested now with volblocksizes of 32k and 8k (the default).
This settings last longer but both produce bluescreens during random write.
Driver version was 1.185.