Description of problem: Host: F11 x86_64 VM: Centos 5.3 x86_64 After creating a RAID5, instantly creating a lvm physical volume with /dev/md0 causes VM to crash. /var/log/dmesg: [...] kvm: 3163: cpu0 unimplemented perfctr wrmsr: 0xc0010004 data 0x0 kvm: 3163: cpu0 unimplemented perfctr wrmsr: 0xc0010000 data 0x130076 kvm: 3163: cpu0 unimplemented perfctr wrmsr: 0xc0010004 data 0xffffffffffdce732 kvm: 3163: cpu0 unimplemented perfctr wrmsr: 0xc0010000 data 0x530076 kvm: 3163: cpu1 unimplemented perfctr wrmsr: 0xc0010004 data 0x0 kvm: 3163: cpu1 unimplemented perfctr wrmsr: 0xc0010000 data 0x130076 kvm: 3163: cpu1 unimplemented perfctr wrmsr: 0xc0010004 data 0xffffffffffdce732 kvm: 3163: cpu1 unimplemented perfctr wrmsr: 0xc0010000 data 0x530076 kvm: 3163: cpu2 unimplemented perfctr wrmsr: 0xc0010004 data 0x0 kvm: 3163: cpu2 unimplemented perfctr wrmsr: 0xc0010000 data 0x130076 __ratelimit: 2 callbacks suppressed qemu-kvm[3163]: segfault at 0 ip 00000000004735c5 sp 00007fff54d0f680 error 4 in qemu-kvm[400000+1da000] [...] Version-Release number of selected component (if applicable): [~]# uname -r 2.6.30.5-43.fc11.x86_64 [~]# rpm -qi qemu-kvm Name : qemu-kvm Relocations: (not relocatable) Version : 0.10.6 Vendor: Fedora Project Release : 4.fc11 Build Date: Fr 04 Sep 2009 12:50:38 CEST [...] How reproducible: Everytime overwriting /dev/md0 over synced area. Steps to Reproduce: 1. mdadm -C -l 5 /dev/md0 ... 2. pvcreate /dev/md0 3. Actual results: VM crashes. Expected results: PV should be created. Additional info: N/A
Thanks for the report Could you include details of the guest configuration etc.? See: https://fedoraproject.org/wiki/Reporting_virtualization_bugs (In reply to comment #0) > kvm: 3163: cpu0 unimplemented perfctr wrmsr: 0xc0010004 data 0x0 these are just warnings, I doubt they're the cause of your crash > qemu-kvm[3163]: segfault at 0 ip 00000000004735c5 sp 00007fff54d0f680 error 4 > in qemu-kvm[400000+1da000] To figure out what caused this segfault, we need to get a stack trace best way is probably to install qemu-debuginfo, start the guest, do gdb qemu-kvm $pid_of_process, 'cont' and trigger the crash; then do 'thread apply all bt full'
No response to needinfo since 2009-10-01, closing