Bug 848955

Note that this worked fine for the same VM with only a different qemu version (both an older qemu-kvm-rhev and qemu-kvm).
All else remained equal.

Hi, Liron

kvm qe can not reproduce this issue via qemu-kvm command directly even lots hotunplug/hotplug repeatly, would you please answer the following questions?

1. reproducible rate ? 100%?
2. please attach the libvirt/qemu/vdsm logs
3. please provide the libvirt and vdsm version.

Additional infos
1. I also have a general testing via rhelm, can not reproduce.
#qemu-kvm-rhev-0.12.1.2-2.304.el6.x86_64
#libvirt-0.9.10-21.el6_3.3.x86_64
#vdsm-4.9.6-27.0.el6_3.x86_64


2. qemu-kvm reproduce steps 
1.start a rhel6.2 guest
/usr/libexec/qemu-kvm -M rhel6.2.0 -enable-kvm -m 2048 -smp
4,sockets=2,cores=2,threads=1 -name rhel6.3 -uuid
8842aa9e-2d20-4540-8557-4d04752a28d7 -drive
file=rhel6.2.qcow2,if=none,id=drive-virtio-disk0,format=qcow2,serial=f82002eb-520c-469b-90c2-663277e90437,cache=none,werror=stop,rerror=stop,aio=native
-device
virtio-blk-pci,scsi=off,bus=pci.0,addr=0x5,drive=drive-virtio-disk0,id=virtio-disk0,bootindex=1
-spice disable-ticketing,port=5913 -vga qxl -drive
file=first-disk.qcow2,if=none,id=drive-virtio-disk1,format=qcow2,cache=none,werror=stop,rerror=stop,aio=native
-device
virtio-blk-pci,scsi=off,bus=pci.0,addr=0x6,drive=drive-virtio-disk1,id=virtio-disk1
-monitor stdio -device sga -chardev
socket,id=serial0,path=/var/test1,server,nowait -device
isa-serial,chardev=serial0 -monitor unix:/var/mm,server,nowait -qmp
tcp:0:6666,server,nowait

2. hot plug three virtio-disks with script
    i=1
while [ $i -lt 32 ]
do
echo "__com.redhat_drive_del drive-virtio-disk1" |nc -U /var/mm
echo "device_del virtio-disk1" |nc -U /var/mm
echo "__com.redhat_drive_del drive-virtio-disk2" |nc -U /var/mm
echo "device_del virtio-disk2" |nc -U /var/mm
echo "__com.redhat_drive_del drive-virtio-disk3" |nc -U /var/mm
echo "device_del virtio-disk3" |nc -U /var/mm
#sleep 3
echo "__com.redhat_drive_add
file=first-disk.qcow2,id=drive-virtio-disk1,format=qcow2,cache=none" |nc
-U /var/mm
echo "device_add
virtio-blk-pci,scsi=off,bus=pci.0,addr=0x6,drive=drive-virtio-disk1,id=virtio-disk1"|nc
-U /var/mm

sleep 2
echo "__com.redhat_drive_add
file=second-disk.qcow2,id=drive-virtio-disk2,format=qcow2,cache=none"
|nc -U /var/mm
echo "device_add
virtio-blk-pci,scsi=off,bus=pci.0,addr=0x7,drive=drive-virtio-disk2,id=virtio-disk2"|nc
-U /var/mm

sleep 2
echo "__com.redhat_drive_add
file=third-disk.qcow2,id=drive-virtio-disk3,format=qcow2,cache=none" |nc
-U /var/mm
echo "device_add
virtio-blk-pci,scsi=off,bus=pci.0,addr=0x8,drive=drive-virtio-disk3,id=virtio-disk3"|nc
-U /var/mm

i=$(($i+1))
sleep 1
done

3. monitor terminal and system log
   monitor qmp:
   telnet 10.66.104.54 6666
   {"execute": "qmp_capabilities"}

   monitor serial output
   nc -U /var/test1

    monitor syslog in guest
    tail -f /var/log/messages

Created attachment 606173 [details]
vdsm log

vdsm version :
vdsm-4.10.0-0.313.git8bedc7e.el6.x86_64

libvirt version:
libvirt-0.9.10-21.el6.x86_64


futher information - i tried this on two hosts, in one of them when i reinstalled the newer version (qemu-kvm-rhev-0.12.1.2-2.304.el6.x86_64) it did work after earlier it didn't - so i don't know what causes this error, but it seems to be some inconsistent behaviour that it's worth to check.

in one host it didn't work with either version of qemu-kvm-rhev (the attached log is from that host).

FYI
Bug 807023 - libvirt does not check for successful device_del

Please provide /var/log/libvirt/qemu/<guest-name>.log for a failed test.