2016-10-19 15:09 GMT+03:00 Guido Günther <a...@sigxcpu.org>: > control: severity -1 normal > control: tags -1 +help > > Hi, > On Wed, Oct 19, 2016 at 12:00:26PM +0000, Mathieu Tarral wrote: >> Package: libvirt-bin >> Version: 1.2.9-9+deb8u3 >> Severity: important >> >> Dear Maintainer, >> >> I upgraded my QEMU package to 2.6+dfsg-3.1~bpo8+1, available from Jessie >> backports. >> However, when i tried to use it, running my virtual machines with virsh, >> i got this error : >> >> error: internal error: process exited while connecting to monitor: >> 2016-10-19T11:50:54.573141Z >> qemu-system-x86_64: -drive if=none,id=drive-scsi0-0-1,readonly=on,format=raw: >> Can't use 'raw' as a block driver for the protocol level >> >> Discussing the issue with maintainers on IRC, we understood that the command >> line >> generated by libvirt is wrong. >> >> In fact, the bug has been fixed in libvirt 1.3.1 >> https://www.spinics.net/linux/fedora/libvir/msg138220.html >> >> Could you backport libvirt-bin v1.3.1 at least in Jessie ? > > Lets rather fix the version in Jessie and let all users benefit from the > fix. Could you check that the patch mentioned in your link > > d7db33bfe978c89e1302609ac91e65be3d49379f > > works with the QEMU in Jessie and with the backported one when applied > to 1.2.9? > > Cheers, > -- Guido
One of the maintainer on IRC tested the QEMU bug against multiple libvirt versions, and that's how he found that libvirt v1.3.3 was working. Again, only QEMU v2.6 from backport is affected. I'm not in a good configuration to test the libvirt patch as i'm working on Fedora, and i don't have full access to the Debian box where i found the bug. Could you take care of it ? Thanks !