[virt-tools-list] [virt-manager PATCH 4/6] devices: disk: Use virtio-scsi for CDROMs where supported

[Cole Robinson]

On 3/8/19 8:04 AM, Andrea Bolognani wrote:
> On Wed, 2019-03-06 at 14:37 -0500, Cole Robinson wrote:
>> On 3/4/19 11:11 AM, Andrea Bolognani wrote:
>>> We already use virtio-blk for regular disks whenever possible,
>>> and there's no good reason not to do the same with virtio-scsi
>>> when dealing with CDROMs instead of artificially limiting its
>>> use to s390x and ppc64/pseries guests.
>>>
>>> Signed-off-by: Andrea Bolognani <abologna at redhat.com>
>>> ---
>>>  .../compare/virt-install-kvm-session-defaults.xml           | 6 ++++--
>>>  tests/cli-test-xml/compare/virt-install-location-iso.xml    | 6 ++++--
>>>  virtinst/devices/disk.py                                    | 4 +---
>>>  3 files changed, 9 insertions(+), 7 deletions(-)
>>
>> This one makes me a bit uncomfortable because we are changing a long
>> held x86 default... but whatever, I'm willing to give it a spin and see
>> if it causes problems.
> 
> I'm afraid you might have been right about this one :(
> 
> I've just tried installing a bunch of guests from virtio-scsi CDROM
> and while Fedora, CentOS and Ubuntu all manage just fine, Debian
> can't locate its own installation source and leaves you hanging.
> 
> Interestingly, Debian has no problem installing to a disk connected
> to that same virtio-scsi controller...
> 
> I'll perform some more installation tests with various distributions
> on non-x86 architectures to see whether it's broken there as well,
> but in the meantime this commit should be reverted. Sorry :(
> 

No worries I committed it so I'm on the hook too. Thanks for testing
I'm glad it didn't hit a release. I've reverted it now

- Cole