Red Hat Bugzilla – Bug 559163
migration failed host 5.5 with -M rhel5.5.0 to host 5.5 with -M rhel5.5.0.
Last modified: 2010-03-30 03:54:22 EDT
Description of problem: Version-Release number of selected component (if applicable): host 5.5.0 ---kvm-83-152 (host kernel- 2.6.18-185.el5) How reproducible: 100% Steps to Reproduce: 1.start guest in host 5.5 with -M rhel5.5.0 # /usr/libexec/qemu-kvm -drive file=win28k-r2-ide-147.qcow2,if=ide -no-hpet -rtc-td-hack -usbdevice tablet -startdate now -smp 2 -m 2G -net nic,model=e1000,macaddr=20:20:20:11:23:12,vlan=0 -net tap,vlan=0,script=/etc/qemu-ifup -cpu qemu64,+sse2 -name r2 -monitor stdio -vnc :6 -M rhel5.5.0 2.start listening mode in local host 5.5 with -M rhel5.5.0 # /usr/libexec/qemu-kvm -drive file=win28k-r2-ide-147.qcow2,if=ide -no-hpet -rtc-td-hack -usbdevice tablet -startdate now -smp 2 -m 2G -net nic,model=e1000,macaddr=20:20:20:11:23:12,vlan=0 -net tap,vlan=0,script=/etc/qemu-ifup -cpu qemu64,+sse2 -name r2 -monitor stdio -vnc :6 -M rhel5.5.0 -incoming tcp:0:5889 3.start migration Actual results: In dest host qemu monitor: (qemu) Unknown savevm section type 255 load of migration failed Expected results: Additional info: test with another linux guest, guest reboot in dest host after migration. cmd: /usr/libexec/qemu-kvm -drive file=RHEL-Server-5.4-64-virtio.qcow2,if=virtio,boot=on -no-hpet -rtc-td-hack -usbdevice tablet -startdate now -smp 2 -m 2G -net nic,model=virtio,macaddr=20:20:20:11:23:16,vlan=0 -net tap,vlan=0,script=/etc/qemu-ifup -cpu qemu64,+sse2 -name 64 -monitor stdio -vnc :9 -no-kvm-pit-reinjection -M rhel5.5.0
Yesterday patch to fix migration between 5.4.4 <-> 5.5.0 broke migration between 5.5.0 <-> 5.5.0. How did it passed my testing is still a mystery (but that I get confused in too many terminals and tested a wrong binary is a posibility). Problem was that at the start at of cpu_load() we check for the valid versions. I forgot to update that version_id == 9 is a valid version. Just fixed it "forever" and used a right test there. While testing this, I found another bug on ide. How this didn't show up sooner is a bigger mistery(*). Problem here was that ide did this checked for version_id == 3 when it meaned version_id > 2 (to be future proof). I send both for this bugzilla. Eduardo, if you want I can create another bugzilla for the ide one, but I think that the problem here was the combination of the two bugs. I tested this time: 5.5.0 -M 5.4.0 -> 5.5.0 -M 5.4.0 pass 5.5.0 -M 5.4.0 -> 5.5.0 -M 5.4.4 pass 5.5.0 -M 5.4.0 -> 5.5.0 -M 5.5.0 pass 5.5.0 -M 5.4.4 -> 5.5.0 -M 5.4.0 fail as expected 5.5.0 -M 5.4.4 -> 5.5.0 -M 5.4.4 pass 5.5.0 -M 5.4.4 -> 5.5.0 -M 5.5.0 pass 5.5.0 -M 5.5.0 -> 5.5.0 -M 5.4.0 fail as expected 5.5.0 -M 5.5.0 -> 5.5.0 -M 5.4.4 fail as expected 5.5.0 -M 5.5.0 -> 5.5.0 -M 5.5.0 pass 5.4.4 -M 5.4.0 -> 5.5.0 -M 5.4.0 pass 5.4.4 -M 5.4.0 -> 5.5.0 -M 5.4.4 pass 5.4.4 -M 5.4.0 -> 5.5.0 -M 5.5.0 pass 5.4.4 -M 5.4.4 -> 5.5.0 -M 5.4.0 fail as expected 5.4.4 -M 5.4.4 -> 5.5.0 -M 5.4.4 pass 5.4.4 -M 5.4.4 -> 5.5.0 -M 5.5.0 pass 5.5.0 -M 5.4.0 -> 5.4.4 -M 5.4.0 pass 5.5.0 -M 5.4.0 -> 5.4.4 -M 5.4.4 pass 5.5.0 -M 5.4.4 -> 5.4.4 -M 5.4.0 fail as expected 5.5.0 -M 5.4.4 -> 5.4.4 -M 5.4.4 pass 5.5.0 -M 5.5.0 -> 5.4.4 -M 5.4.0 fail as expected 5.5.0 -M 5.5.0 -> 5.4.4 -M 5.4.4 fail as expected (And that is without testing 5.4.0 proper)
PASS for the following scenarios. host 5.5.0 ---kvm-83-154 host 5.4.4 ---kvm-83-105.el5.4.19 | id | source | dest | block | result | | 1 | 5.5.0 with -M rhel5.4.4 | 5.4.4 with -M rhel5.4.4 | ide | pass | | 2 | 5.5.0 with -M rhel5.4.4 | 5.4.4 with -M rhel5.4.4 | virtio | pass | | 3 | 5.4.4 with -M rhel5.4.4 | 5.5.0 with -M rhel5.4.4 | ide | pass | | 4 | 5.4.4 with -M rhel5.4.4 | 5.5.0 with -M rhel5.4.4 | virtio | pass | | 5 | 5.5.0 with -M rhel5.4.0 | 5.4.4 with -M rhel5.4.0 | ide | pass | | 6 | 5.5.0 with -M rhel5.4.0 | 5.4.4 with -M rhel5.4.0 | virtio | pass | | 7 | 5.4.4 with -M rhel5.4.0 | 5.5.0 with -M rhel5.4.0 | ide | pass | | 8 | 5.4.4 with -M rhel5.4.0 | 5.5.0 with -M rhel5.4.0 | virtio | pass | | 9 | 5.5.0 with -M rhel5.5.0 | 5.5.0 with -M rhel5.5.0 | ide | pass | | 10 | 5.5.0 with -M rhel5.5.0 | 5.5.0 with -M rhel5.5.0 | virtio | pass | | 11 | 5.5.0 with -M rhel5.4.4 | 5.5.0 with -M rhel5.4.4 | ide | pass | | 12 | 5.5.0 with -M rhel5.4.4 | 5.5.0 with -M rhel5.4.4 | virtio | pass | | 13 | 5.5.0 with -M rhel5.4.0 | 5.5.0 with -M rhel5.4.0 | ide | pass | | 14 | 5.5.0 with -M rhel5.4.0 | 5.5.0 with -M rhel5.4.0 | virtio | pass | | 15 | 5.5.0 with -M rhel5.4.0 | 5.5.0 with -M rhel5.4.4 | ide | pass | | 16 | 5.5.0 with -M rhel5.4.0 | 5.5.0 with -M rhel5.4.4 | virtio | pass | | 17 | 5.5.0 with -M rhel5.4.0 | 5.5.0 with -M rhel5.5.0 | ide | pass | | 18 | 5.5.0 with -M rhel5.4.0 | 5.5.0 with -M rhel5.5.0 | virtio | pass | | 19 | 5.5.0 with -M rhel5.4.4 | 5.5.0 with -M rhel5.5.0 | ide | pass | | 20 | 5.5.0 with -M rhel5.4.4 | 5.5.0 with -M rhel5.5.0 | virtio | pass | | 21 | 5.4.4 with -M rhel5.4.0 | 5.5.0 with -M rhel5.4.4 | ide | pass | | 22 | 5.4.4 with -M rhel5.4.0 | 5.5.0 with -M rhel5.4.4 | virtio | pass | | 23 | 5.4.4 with -M rhel5.4.0 | 5.5.0 with -M rhel5.5.0 | ide | pass | | 24 | 5.4.4 with -M rhel5.4.0 | 5.5.0 with -M rhel5.5.0 | virtio | pass | | 25 | 5.4.4 with -M rhel5.4.4 | 5.5.0 with -M rhel5.5.0 | ide | pass | | 26 | 5.4.4 with -M rhel5.4.4 | 5.5.0 with -M rhel5.5.0 | virtio | pass | | 27 | 5.5.0 with -M rhel5.4.0 | 5.4.4 with -M rhel5.4.4 | ide | pass | | 28 | 5.5.0 with -M rhel5.4.0 | 5.4.4 with -M rhel5.4.4 | virtio | pass |
Were there other changes, for example in virtio-net, that should be tested as well?
Add nic info for comment#4: | id | source | dest | block | nic | result | | 1 | 5.5.0 with -M rhel5.4.4 | 5.4.4 with -M rhel5.4.4 | ide | rtl8139 | pass | | 2 | 5.5.0 with -M rhel5.4.4 | 5.4.4 with -M rhel5.4.4 | virtio | rtl8139 | pass | | 3 | 5.4.4 with -M rhel5.4.4 | 5.5.0 with -M rhel5.4.4 | ide | rtl8139 | pass | | 4 | 5.4.4 with -M rhel5.4.4 | 5.5.0 with -M rhel5.4.4 | virtio | rtl8139 | pass | | 5 | 5.5.0 with -M rhel5.4.0 | 5.4.4 with -M rhel5.4.0 | ide | rtl8139 | pass | | 6 | 5.5.0 with -M rhel5.4.0 | 5.4.4 with -M rhel5.4.0 | virtio | rtl8139 | pass | | 7 | 5.4.4 with -M rhel5.4.0 | 5.5.0 with -M rhel5.4.0 | ide | rtl8139 | pass | | 8 | 5.4.4 with -M rhel5.4.0 | 5.5.0 with -M rhel5.4.0 | virtio | rtl8139 | pass | | 9 | 5.5.0 with -M rhel5.5.0 | 5.5.0 with -M rhel5.5.0 | ide | rtl8139 | pass | | 10 | 5.5.0 with -M rhel5.5.0 | 5.5.0 with -M rhel5.5.0 | virtio | rtl8139 | pass | | 11 | 5.5.0 with -M rhel5.4.4 | 5.5.0 with -M rhel5.4.4 | ide | rtl8139 | pass | | 12 | 5.5.0 with -M rhel5.4.4 | 5.5.0 with -M rhel5.4.4 | virtio | rtl8139 | pass | | 13 | 5.5.0 with -M rhel5.4.0 | 5.5.0 with -M rhel5.4.0 | ide | rtl8139 | pass | | 14 | 5.5.0 with -M rhel5.4.0 | 5.5.0 with -M rhel5.4.0 | virtio | rtl8139 | pass | | 15 | 5.5.0 with -M rhel5.4.0 | 5.5.0 with -M rhel5.4.4 | ide | rtl8139 | pass | | 16 | 5.5.0 with -M rhel5.4.0 | 5.5.0 with -M rhel5.4.4 | virtio | virtio | pass | | 17 | 5.5.0 with -M rhel5.4.0 | 5.5.0 with -M rhel5.5.0 | ide | rtl8139 | pass | | 18 | 5.5.0 with -M rhel5.4.0 | 5.5.0 with -M rhel5.5.0 | virtio | virtio | pass | | 19 | 5.5.0 with -M rhel5.4.4 | 5.5.0 with -M rhel5.5.0 | ide | rtl8139 | pass | | 20 | 5.5.0 with -M rhel5.4.4 | 5.5.0 with -M rhel5.5.0 | virtio | virtio | pass | | 21 | 5.4.4 with -M rhel5.4.0 | 5.5.0 with -M rhel5.4.4 | ide | rtl8139 | pass | | 22 | 5.4.4 with -M rhel5.4.0 | 5.5.0 with -M rhel5.4.4 | virtio | virtio | pass | | 23 | 5.4.4 with -M rhel5.4.0 | 5.5.0 with -M rhel5.5.0 | ide | rtl8139 | pass | | 24 | 5.4.4 with -M rhel5.4.0 | 5.5.0 with -M rhel5.5.0 | virtio | virtio | pass | | 25 | 5.4.4 with -M rhel5.4.4 | 5.5.0 with -M rhel5.5.0 | ide | rtl8139 | pass | | 26 | 5.4.4 with -M rhel5.4.4 | 5.5.0 with -M rhel5.5.0 | virtio | virtio | pass | | 27 | 5.5.0 with -M rhel5.4.0 | 5.4.4 with -M rhel5.4.4 | ide | rtl8139 | pass | | 28 | 5.5.0 with -M rhel5.4.0 | 5.4.4 with -M rhel5.4.4 | virtio | virtio | pass |
An advisory has been issued which should help the problem described in this bug report. This report is therefore being closed with a resolution of ERRATA. For more information on therefore solution and/or where to find the updated files, please follow the link below. You may reopen this bug report if the solution does not work for you. http://rhn.redhat.com/errata/RHSA-2010-0271.html