Bug#731815: virsh: Unable to start any VM with custom cpu model
Hi, I had exactly the same problem with latest testing from today. Just delete the old /var/cache/libvirt/qemu/capabilities and do /etc/init.d/libvirtd restart solves all problems. Thx to Ariel Malves hf, Joerg -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#731815: [Pkg-libvirt-maintainers] Bug#731815: virsh: Unable to start any VM with custom cpu model
I did a fresh install from beta2 iso: http://cdimage.debian.org/cdimage/jessie_di_beta_2/amd64/iso-cd/debian-jessie-DI-b2-amd64-lxde-CD-1.iso Following steps to reproduce: After fresh install the /var/cache/libvirt/qemu/capabilities folder does not exist. Installed libvirt-bin and virtinst packages (without apt-get update). Then i see about 20 xml files in cache folder. virt-install --name test --memory 1024 --nodisk --pxe complains that default network is not running virsh net-start default and when i now run virt-install again, it gives me again error of CPU. I didnt touch qemu.conf and libvirtd.conf yet. couple of packages were installed from ISO, not sure which ones. apt-get update and upgrade shows that everything with libvirt is uptodate (1.2.9-3, python-libvirt 1.2.9-1). Am 29.10.2014 06:19, schrieb Guido Günther: That points to libvit no refreshing capabilities properly during the upgrade from wheezy to jessie. It should detect the changed qemu and do so. Can you add the information what versions (at least roughly) of qemu and libvirt you were ugrading from. The versions from stable or stable-backports? Cheers, -- Guido -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#731815: virsh: Unable to start any VM with custom cpu model
Hi, I faced a similar problem when upgrading libvirt/qemu packages from wheezy to wheezy-backports. Try to delete /var/cache/libvirt/qemu/capabilities and restart libvirt-bin. []'s Ariel On 10/27/2014 02:38 PM, Thomas Rechberger wrote: Same problem after changing from wheezy to jessie. virt-install --name test --memory 1024 --nodisk --pxe --cpu host ERRORinternal error: Cannot find suitable CPU model for given data libvirtd log: Preferred CPU model Opteron_G3 not allowed by hypervisor; closest supported model will be used has it anything to do with wrong detected cpu flags? I am using AMD Kabini cpu and Opteron G3 map is used both under wheezy and jessie. virsh capabilities /tmp/capabilities.xml root@kvm:/# virsh cpu-compare /tmp/capabilities.xml CPU described in /tmp/capabilities.xml is identical to host CPU gir1.2-libvirt-glib-1.0:amd64/jessie 0.1.7-2.1 uptodate libvirt-bin:amd64/jessie 1.2.9-3 uptodate libvirt-clients:amd64/jessie 1.2.9-3 uptodate libvirt-daemon:amd64/jessie 1.2.9-3 uptodate libvirt-daemon-system:amd64/jessie 1.2.9-3 uptodate libvirt-glib-1.0-0:amd64/jessie 0.1.7-2.1 uptodate libvirt0:amd64/jessie 1.2.9-3 uptodate cat /proc/cpuinfo processor : 0 vendor_id : AuthenticAMD cpu family : 22 model : 0 model name : AMD Sempron(tm) 3850 APU with Radeon(tm) R3 stepping : 1 flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc rep_good nopl nonstop_tsc extd_apicid aperfmperf eagerfpu pni pclmulqdq monitor ssse3 cx16 sse4_1 sse4_2 movbe popcnt aes xsave avx f16c lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt topoext perfctr_nb perfctr_l2 arat xsaveopt hw_pstate proc_feedback npt lbrv svm_lock nrip_save tsc_scale flushbyasid decodeassists pausefilter pfthreshold bmi1 virsh -r capabilities shows me additional invtsc osxsave feature which the older version didnt had. Here was mentioned a kernel update fixed it: https://bbs.archlinux.org/viewtopic.php?id=182142 What it has to do with kernel version? -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#731815: virsh: Unable to start any VM with custom cpu model
That is totally working for me, thank you so much for the find. Am 28.10.2014 18:33, schrieb Ariel Malves: Hi, I faced a similar problem when upgrading libvirt/qemu packages from wheezy to wheezy-backports. Try to delete /var/cache/libvirt/qemu/capabilities and restart libvirt-bin. []'s Ariel On 10/27/2014 02:38 PM, Thomas Rechberger wrote: Same problem after changing from wheezy to jessie. virt-install --name test --memory 1024 --nodisk --pxe --cpu host ERRORinternal error: Cannot find suitable CPU model for given data libvirtd log: Preferred CPU model Opteron_G3 not allowed by hypervisor; closest supported model will be used has it anything to do with wrong detected cpu flags? I am using AMD Kabini cpu and Opteron G3 map is used both under wheezy and jessie. virsh capabilities /tmp/capabilities.xml root@kvm:/# virsh cpu-compare /tmp/capabilities.xml CPU described in /tmp/capabilities.xml is identical to host CPU gir1.2-libvirt-glib-1.0:amd64/jessie 0.1.7-2.1 uptodate libvirt-bin:amd64/jessie 1.2.9-3 uptodate libvirt-clients:amd64/jessie 1.2.9-3 uptodate libvirt-daemon:amd64/jessie 1.2.9-3 uptodate libvirt-daemon-system:amd64/jessie 1.2.9-3 uptodate libvirt-glib-1.0-0:amd64/jessie 0.1.7-2.1 uptodate libvirt0:amd64/jessie 1.2.9-3 uptodate cat /proc/cpuinfo processor : 0 vendor_id : AuthenticAMD cpu family : 22 model : 0 model name : AMD Sempron(tm) 3850 APU with Radeon(tm) R3 stepping : 1 flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc rep_good nopl nonstop_tsc extd_apicid aperfmperf eagerfpu pni pclmulqdq monitor ssse3 cx16 sse4_1 sse4_2 movbe popcnt aes xsave avx f16c lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt topoext perfctr_nb perfctr_l2 arat xsaveopt hw_pstate proc_feedback npt lbrv svm_lock nrip_save tsc_scale flushbyasid decodeassists pausefilter pfthreshold bmi1 virsh -r capabilities shows me additional invtsc osxsave feature which the older version didnt had. Here was mentioned a kernel update fixed it: https://bbs.archlinux.org/viewtopic.php?id=182142 What it has to do with kernel version? -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#731815: [Pkg-libvirt-maintainers] Bug#731815: virsh: Unable to start any VM with custom cpu model
On Tue, Oct 28, 2014 at 07:36:40PM +0100, Thomas Rechberger wrote: That is totally working for me, thank you so much for the find. That points to libvit no refreshing capabilities properly during the upgrade from wheezy to jessie. It should detect the changed qemu and do so. Can you add the information what versions (at least roughly) of qemu and libvirt you were ugrading from. The versions from stable or stable-backports? Cheers, -- Guido -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#731815: virsh: Unable to start any VM with custom cpu model
Same problem after changing from wheezy to jessie. virt-install --name test --memory 1024 --nodisk --pxe --cpu host ERRORinternal error: Cannot find suitable CPU model for given data libvirtd log: Preferred CPU model Opteron_G3 not allowed by hypervisor; closest supported model will be used has it anything to do with wrong detected cpu flags? I am using AMD Kabini cpu and Opteron G3 map is used both under wheezy and jessie. virsh capabilities /tmp/capabilities.xml root@kvm:/# virsh cpu-compare /tmp/capabilities.xml CPU described in /tmp/capabilities.xml is identical to host CPU gir1.2-libvirt-glib-1.0:amd64/jessie 0.1.7-2.1 uptodate libvirt-bin:amd64/jessie 1.2.9-3 uptodate libvirt-clients:amd64/jessie 1.2.9-3 uptodate libvirt-daemon:amd64/jessie 1.2.9-3 uptodate libvirt-daemon-system:amd64/jessie 1.2.9-3 uptodate libvirt-glib-1.0-0:amd64/jessie 0.1.7-2.1 uptodate libvirt0:amd64/jessie 1.2.9-3 uptodate cat /proc/cpuinfo processor : 0 vendor_id : AuthenticAMD cpu family : 22 model : 0 model name : AMD Sempron(tm) 3850 APU with Radeon(tm) R3 stepping : 1 flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc rep_good nopl nonstop_tsc extd_apicid aperfmperf eagerfpu pni pclmulqdq monitor ssse3 cx16 sse4_1 sse4_2 movbe popcnt aes xsave avx f16c lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt topoext perfctr_nb perfctr_l2 arat xsaveopt hw_pstate proc_feedback npt lbrv svm_lock nrip_save tsc_scale flushbyasid decodeassists pausefilter pfthreshold bmi1 virsh -r capabilities shows me additional invtsc osxsave feature which the older version didnt had. Here was mentioned a kernel update fixed it: https://bbs.archlinux.org/viewtopic.php?id=182142 What it has to do with kernel version? -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#731815: virsh: Unable to start any VM with custom cpu model
Dear Maintainer, Just to confirm the bug: I'm running Debian stable, with libvirt, qemu and kvm from wheezy-backports (libvirt 1.2.1, API QEMU 1.2.1, hypervisor QEMU 1.7.1), in a HP ProLiant G7 server, with Xeon E5649 processor. I can't start a domain, through virsh start, with any type of custom cpu model, even 486. E.g.: cpu mode='custom' match='exact' model fallback='allow'486/model /cpu virsh start DomTest1 output: error: Failed to start domain DomTest1 error: internal error: Cannot find suitable CPU model for given data /var/log/libvirt/libvirtd.log related contents: 2014-04-25 13:26:48.174+: 9968: warning : x86Decode:1511 : Preferred CPU model 486 not allowed by hypervisor; closest supported model will be used 2014-04-25 13:26:48.174+: 9968: error : x86Decode:1564 : internal error: Cannot find suitable CPU model for given data Running qemu-system-x86_64 by hand do not show any errors (qemu-system-x86_64 -enable-kvm -name DomTest1 -S -machine pc-i440fx-1.7,accel=kvm,usb=off -cpu 486 -m 1024.) Please, is there any workaround for this bug? Ariel -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org