Red Hat Bugzilla – Bug 1358728
Migration failed when the secondary video devices have different ram/vram sizes.
Last modified: 2016-11-03 14:50:52 EDT
Description of problem: Guest has three video devices, and the vram or ram size of the second video and the third video are not equal. Start the guest, and migration will fail: # virsh migrate rhel7.2 qemu+ssh://hp-dl385g7-06.lab.eng.pek2.redhat.com/system --live --verbose error: internal error: qemu unexpectedly closed the monitor: 2016-07-20T09:17:43.220520Z qemu-kvm: Length mismatch: 0000:00:09.0/qxl.vram: 0x2000000 in != 0x4000000: Invalid argument 2016-07-20T09:17:43.221314Z qemu-kvm: error while loading state for instance 0x0 of device 'ram' 2016-07-20T09:17:43.222751Z qemu-kvm: load of migration failed: Invalid argument Version-Release number of selected component: libvirt-2.0.0-2.el7.x86_64 How reproducible: 100% Steps to Reproduce: 1. Prepare a guest with three videos, and set different ram or vram size for the second video and the third video: # virsh dumpxml rhel7.2 --inactive ... <video> <model type='qxl' ram='1048576' vram='16384' vgamem='16384' heads='1' primary='yes'/> <address type='pci' domain='0x0000' bus='0x00' slot='0x02' function='0x0'/> </video> <video> <model type='qxl' ram='65536' vram='**32768**' vgamem='8192' heads='3'/> <address type='pci' domain='0x0000' bus='0x00' slot='0x09' function='0x0'/> </video> <video> <model type='qxl' ram='65536' vram='**65536**' vgamem='8192' heads='7'/> <address type='pci' domain='0x0000' bus='0x00' slot='0x0b' function='0x0'/> </video> ... 2.Start guest # virsh start rhel7.2 3.Check live vram/ram size, find that the value of the second video devices becomes 65536: # virsh dumpxml rhel7.2 ... <video> <model type='qxl' ram='524288' vram='16384' vgamem='16384' primary='yes'/> <alias name='video0'/> <address type='pci' domain='0x0000' bus='0x00' slot='0x02' function='0x0'/> </video> <video> <model type='qxl' ram='65536' vram='65536' vgamem='8192'/> <alias name='video1'/> <address type='pci' domain='0x0000' bus='0x00' slot='0x09' function='0x0'/> </video> <video> <model type='qxl' ram='65536' vram='65536' vgamem='8192'/> <alias name='video2'/> <address type='pci' domain='0x0000' bus='0x00' slot='0x0b' function='0x0'/> </video> ... 4.Do migration and failed. Actual results: Migration failed Expected results: Migration can succeed Additional info: 1) The qemu command line on source host: -device qxl-vga,id=video0,ram_size=1073741824,vram_size=16777216,vram64_size_mb=0,vgamem_mb=16,bus=pci.0,addr=0x2 -device qxl,id=video1,ram_size=67108864,vram_size=33554432,vram64_size_mb=0,vgamem_mb=8,bus=pci.0,addr=0x9 -device qxl,id=video2,ram_size=67108864,vram_size=67108864,vram64_size_mb=0,vgamem_mb=8,bus=pci.0,addr=0xb 2) The qemu command line on target host: -device qxl-vga,id=video0,ram_size=536870912,vram_size=16777216,vram64_size_mb=0,vgamem_mb=16,bus=pci.0,addr=0x2 -device qxl,id=video1,ram_size=67108864,vram_size=67108864,vram64_size_mb=0,vgamem_mb=8,bus=pci.0,addr=0x9 -device qxl,id=video2,ram_size=67108864,vram_size=67108864,vram64_size_mb=0,vgamem_mb=8,bus=pci.0,addr=0xb 3) I tried several times, and found that the vram/ram size of the second video device is always adjusted to the vram/ram value of the third video device
Upstream commit: commit 4a585a884fc8cfd11f6206cd6f1a7db824395e86 Author: Pavel Hrdina <phrdina@redhat.com> Date: Mon Jul 25 10:47:00 2016 +0200 qemu_monitor_json: add support to search QOM device path by device alias
Can reproduce this BZ with libvirt-2.0.0-2.el7.x86_64. Verified pass with libvirt-2.0.0-5.el7.x86_64. Test steps: 1. Prepare a guest with three videos, and set different ram or vram size for the second video and the third video: # virsh dumpxml rhel7.2 --inactive ... <video> <model type='qxl' ram='1048576' vram='16384' vgamem='16384' heads='1' primary='yes'/> <address type='pci' domain='0x0000' bus='0x00' slot='0x02' function='0x0'/> </video> <video> <model type='qxl' ram='65536' vram='**32768**' vgamem='8192' heads='3'/> <address type='pci' domain='0x0000' bus='0x00' slot='0x09' function='0x0'/> </video> <video> <model type='qxl' ram='65536' vram='**65536**' vgamem='8192' heads='7'/> <address type='pci' domain='0x0000' bus='0x00' slot='0x0b' function='0x0'/> </video> ... 2.Start guest # virsh start rhel7.2 3.Check live vram/ram size, find all the value of video devices are not changes: # virsh dumpxml rhel7.2 ... <video> <model type='qxl' ram='524288' vram='16384' vgamem='16384' primary='yes'/> <alias name='video0'/> <address type='pci' domain='0x0000' bus='0x00' slot='0x02' function='0x0'/> </video> <video> <model type='qxl' ram='65536' vram='**32768**' vgamem='8192'/> <alias name='video1'/> <address type='pci' domain='0x0000' bus='0x00' slot='0x09' function='0x0'/> </video> <video> <model type='qxl' ram='65536' vram='**65536**' vgamem='8192'/> <alias name='video2'/> <address type='pci' domain='0x0000' bus='0x00' slot='0x0b' function='0x0'/> </video> ... 4.Do migration,migration complete correctly and guest works well after migration: #virsh migrate rhel7.2 qemu+ssh://10.66.144.76/system --live --verbose Migration: [100 %] 5.Also try four video devices with different ram/vram size, all the value of video devices are not changed when start the guest.
Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory, and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. https://rhn.redhat.com/errata/RHSA-2016-2577.html