<danpb> dvd: with virtio devices hot-unplug results in TWO DEVICE_DELETED events
<dvd> danpb, this is the bz I was preparing: http://pastebin.test.redhat.com/1021567
<dvd> oh
<danpb> libvirt ignores one of them (the one you show above without any 'device' field
<sean-k-mooney> danpb: we are seeing this upstream too i think so we likely will have to work with canonical to get that backportaded and do it for centos/fedora too
<danpb> QEMU fails to send the second event thuat we actually need
<danpb> libvirt 7.9.0 / 7.10.0, when used in combination with qemu 6.2.0
<danpb> there is a libvirt 8.0.0 build being done today that will fix this by reverting to non-json usage
<danpb> until my qemu fix works its way in
One of them is because qemu doesn’t send the 2nd device_deleted event: /lists. gnu.org/ archive/ html/qemu- devel/2022- 01/msg00542. html
https:/
<danpb> dvd: with virtio devices hot-unplug results in TWO DEVICE_DELETED events pastebin. test.redhat. com/1021567
<dvd> danpb, this is the bz I was preparing: http://
<dvd> oh
<danpb> libvirt ignores one of them (the one you show above without any 'device' field
<sean-k-mooney> danpb: we are seeing this upstream too i think so we likely will have to work with canonical to get that backportaded and do it for centos/fedora too
<danpb> QEMU fails to send the second event thuat we actually need
<danpb> libvirt 7.9.0 / 7.10.0, when used in combination with qemu 6.2.0
<danpb> there is a libvirt 8.0.0 build being done today that will fix this by reverting to non-json usage
<danpb> until my qemu fix works its way in
More details under the "Bug fix" section here: https:/ /libvirt. org/news. html#v8- 0-0-2022- 01-14