Re: Errors with multiple VMPVSCSI devices

[John Call]

Oops, should have also included my versions:
$ virt-install --version
3.2.0

$ virsh -v
6.6.0

$ qemu-system-x86_64 -version
QEMU emulator version 5.1.0 (qemu-5.1.0-9.fc33)
Copyright (c) 2003-2020 Fabrice Bellard and the QEMU Project developers

$ grep PRETTY /etc/os-release
PRETTY_NAME="Fedora 33 (Thirty Three)"

$ cat /proc/version
Linux version 5.11.17-200.fc33.x86_64
(mockbuild@bkernel01.iad2.fedoraproject.org) (gcc (GCC) 10.3.1
20210422 (Red Hat 10.3.1-1), GNU ld version 2.35-18.fc33) #1 SMP Wed
Apr 28 17:34:39 UTC 2021

On Wed, May 12, 2021 at 10:14 AM John Call <johnsimcall@gmail.com> wrote:
>
> Hi qemu-discuss,
>
> I recently needed to boot a VM with many disks attached to a VMPVSCSI
> controller. I had a lot of trouble with this. My biggest issue is the
> "bad scsi device lun: 1" error, which I think is a qemu thing. Is this
> the right place to ask if this can be fixed, or do I need to approach
> the `virt-install` or `libvirt` groups?
>
> I use `virt-install` like this[1] to create my VM, but it refuses to
> start with this error:
> ERROR    internal error: qemu unexpectedly closed the monitor:
> 2021-05-12T15:49:51.310324Z qemu-system-x86_64: -device
> scsi-hd,bus=scsi0.0,channel=0,scsi-id=0,lun=1,device_id=00000003,drive=libvirt-20-format,id=scsi0-0-0-1,serial=00000003:
> bad scsi device lun: 1
>
> I can fix this by grabbing the `/usr/bin/qemu-system-x86_64`
> command-line and changing the incrementing LUN IDs to instead become
> incrementing TARGET IDs like this...
> [jcall@jcall-laptop tmp]$ diff -U0 a-pvscsi.sh b-pvscsi.sh
> --- a-pvscsi.sh 2021-05-12 09:10:55.427049902 -0600
> +++ b-pvscsi.sh 2021-05-12 09:10:59.715114940 -0600
> @@ -41 +17 @@
> --device 
> scsi-hd,bus=scsi0.0,channel=0,scsi-id=1,lun=0,device_id=00000003,drive=libvirt-20-format,id=scsi0-0-0-1,serial=00000003
> \
> +-device 
> scsi-hd,bus=scsi0.0,channel=0,scsi-id=1,lun=0,device_id=00000003,drive=libvirt-20-format,id=scsi0-0-1-0,serial=00000003
> \
> @@ -44 +20 @@
> --device 
> scsi-hd,bus=scsi0.0,channel=0,scsi-id=2,lun=0,device_id=00000004,drive=libvirt-19-format,id=scsi0-0-0-2,serial=00000004
> \
> +-device 
> scsi-hd,bus=scsi0.0,channel=0,scsi-id=2,lun=0,device_id=00000004,drive=libvirt-19-format,id=scsi0-0-2-0,serial=00000004
> \
>
> Now my VM boots! I'd love it if this could "just work" in the future.
> Thanks for all your help!
> John
>
> [1] ## virt-install command ##
> sudo virt-install --name pvscsi-test-sh \
>   --memory $((1024*6)) --vcpus 2 --cpu host \
>   --controller scsi,model=vmpvscsi \
>   --import \
>   --disk vol=default/8.2.2.0-disk1.qcow2,bus=ide,serial=00000001,boot.order=1 
> \
>   --disk vol=default/8.2.2.0-disk2.qcow2,bus=scsi,serial=00000002 \
>   --disk vol=default/8.2.2.0-disk3.qcow2,bus=scsi,serial=00000003 \
>   --disk vol=default/8.2.2.0-disk4.qcow2,bus=scsi,serial=00000004 \
>   --network model=e1000,network=default
> ## end virt-install command ##