Bug 1521202
| Summary: | libvirt change the running guest cpu model after upgrade from 7.3 | ||
|---|---|---|---|
| Product: | Red Hat Enterprise Linux 7 | Reporter: | Luyao Huang <lhuang> |
| Component: | libvirt | Assignee: | Jiri Denemark <jdenemar> |
| Status: | CLOSED ERRATA | QA Contact: | Luyao Huang <lhuang> |
| Severity: | medium | Docs Contact: | |
| Priority: | medium | ||
| Version: | 7.5 | CC: | dyuan, jdenemar, jishao, lhuang, lizhu, lmiksik, rbalakri, xuzhang, yalzhang, zpeng |
| Target Milestone: | rc | ||
| Target Release: | --- | ||
| Hardware: | x86_64 | ||
| OS: | Linux | ||
| Whiteboard: | |||
| Fixed In Version: | libvirt-3.9.0-6.el7 | Doc Type: | If docs needed, set a value |
| Doc Text: | Story Points: | --- | |
| Clone Of: | Environment: | ||
| Last Closed: | 2018-04-10 11:00:58 UTC | Type: | Bug |
| Regression: | --- | Mount Type: | --- |
| Documentation: | --- | CRM: | |
| Verified Versions: | Category: | --- | |
| oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |
| Cloudforms Team: | --- | Target Upstream Version: | |
| Embargoed: | |||
| Bug Depends On: | |||
| Bug Blocks: | 1199452 | ||
Patches sent upstream for review: https://www.redhat.com/archives/libvir-list/2017-December/msg00250.html Fixed upstream by
commit 4486dcdb4a386a906807d4eb7de586202e50cd18
Refs: v3.10.0-37-g4486dcdb4a
Author: Jiri Denemark <jdenemar>
AuthorDate: Thu Dec 7 11:21:48 2017 +0100
Commit: Jiri Denemark <jdenemar>
CommitDate: Fri Dec 8 15:50:52 2017 +0100
qemu: Separate fetching CPU definitions from filling qemuCaps
virQEMUCapsProbeQMPCPUDefinitions is now a small wrapper which fills in
qemuCaps with CPU models fetched by virQEMUCapsFetchCPUDefinitions.
Signed-off-by: Jiri Denemark <jdenemar>
commit ce73de441d35e17a29b83b8741e668a6ddc013ba
Refs: v3.10.0-38-gce73de441d
Author: Jiri Denemark <jdenemar>
AuthorDate: Thu Dec 7 11:23:50 2017 +0100
Commit: Jiri Denemark <jdenemar>
CommitDate: Fri Dec 8 15:50:52 2017 +0100
qemu: Make sure host-model uses CPU model supported by QEMU
When reconnecting to a running domain started by old libvirt, which did
not change host-model into a custom CPU definition, we replace the CPU
definition with a specific CPU model from host capabilities. However,
that CPU model may not be supported by the running qemu process. We need
to translate the CPU model to one of the models which libvirt could have
used when starting the domain.
https://bugzilla.redhat.com/show_bug.cgi?id=1521202
Signed-off-by: Jiri Denemark <jdenemar>
Hi jirka, I found that still get failure when migrate to a rhel7.3 host, but the error is different this time (the step is the same with comment 0): 1. the guest xml before upgrade to 7.5: <cpu mode='host-model' match='exact'> <model fallback='forbid'>Opteron_G3</model> <vendor>AMD</vendor> <feature policy='require' name='vme'/> <feature policy='require' name='ht'/> <feature policy='require' name='pclmuldq'/> <feature policy='require' name='ssse3'/> <feature policy='require' name='fma'/> <feature policy='require' name='sse4.1'/> <feature policy='require' name='sse4.2'/> <feature policy='require' name='movbe'/> <feature policy='require' name='aes'/> <feature policy='require' name='xsave'/> <feature policy='require' name='osxsave'/> <feature policy='require' name='avx'/> <feature policy='require' name='f16c'/> <feature policy='require' name='rdrand'/> <feature policy='require' name='arat'/> <feature policy='require' name='fsgsbase'/> <feature policy='require' name='bmi1'/> <feature policy='require' name='avx2'/> <feature policy='require' name='smep'/> <feature policy='require' name='bmi2'/> <feature policy='require' name='rdseed'/> <feature policy='require' name='adx'/> <feature policy='require' name='smap'/> <feature policy='require' name='clflushopt'/> <feature policy='require' name='xsaveopt'/> <feature policy='require' name='xsavec'/> <feature policy='require' name='xgetbv1'/> <feature policy='require' name='mmxext'/> <feature policy='require' name='fxsr_opt'/> <feature policy='require' name='pdpe1gb'/> <feature policy='require' name='cmp_legacy'/> <feature policy='require' name='extapic'/> <feature policy='require' name='cr8legacy'/> <feature policy='require' name='3dnowprefetch'/> <feature policy='require' name='osvw'/> <feature policy='require' name='skinit'/> <feature policy='require' name='wdt'/> <feature policy='require' name='tce'/> <feature policy='require' name='topoext'/> <feature policy='require' name='perfctr_core'/> <feature policy='require' name='perfctr_nb'/> <feature policy='require' name='ibpb'/> 2. guest xml after update to 7.5: # virsh dumpxml vm1 ... <cpu mode='custom' match='exact' check='full'> <model fallback='allow'>Opteron_G5</model> <vendor>AMD</vendor> <feature policy='require' name='vme'/> <feature policy='require' name='x2apic'/> <feature policy='require' name='movbe'/> <feature policy='require' name='rdrand'/> <feature policy='require' name='hypervisor'/> <feature policy='require' name='arat'/> <feature policy='require' name='fsgsbase'/> <feature policy='require' name='bmi1'/> <feature policy='require' name='avx2'/> <feature policy='require' name='smep'/> <feature policy='require' name='bmi2'/> <feature policy='require' name='rdseed'/> <feature policy='require' name='adx'/> <feature policy='require' name='smap'/> <feature policy='require' name='clflushopt'/> <feature policy='require' name='xsaveopt'/> <feature policy='require' name='xsavec'/> <feature policy='require' name='xgetbv1'/> <feature policy='require' name='mmxext'/> <feature policy='require' name='fxsr_opt'/> <feature policy='require' name='cmp_legacy'/> <feature policy='require' name='cr8legacy'/> <feature policy='require' name='osvw'/> <feature policy='require' name='ibpb'/> <feature policy='disable' name='rdtscp'/> <feature policy='disable' name='svm'/> <feature policy='disable' name='xop'/> <feature policy='disable' name='fma4'/> <feature policy='disable' name='tbm'/> <feature policy='disable' name='ht'/> <feature policy='disable' name='osxsave'/> <feature policy='disable' name='extapic'/> <feature policy='disable' name='skinit'/> <feature policy='disable' name='wdt'/> <feature policy='disable' name='tce'/> <feature policy='disable' name='topoext'/> <feature policy='disable' name='perfctr_core'/> <feature policy='disable' name='perfctr_nb'/> <feature policy='disable' name='monitor'/> <feature policy='disable' name='sha-ni'/> ... # virsh dumpxml vm1 --migratable ... <cpu mode='custom' match='exact' check='partial'> <model fallback='allow'>EPYC</model> <vendor>AMD</vendor> <feature policy='require' name='ht'/> <feature policy='require' name='osxsave'/> <feature policy='require' name='cmp_legacy'/> <feature policy='require' name='extapic'/> <feature policy='require' name='skinit'/> <feature policy='require' name='wdt'/> <feature policy='require' name='tce'/> <feature policy='require' name='topoext'/> <feature policy='require' name='perfctr_core'/> <feature policy='require' name='perfctr_nb'/> <feature policy='require' name='ibpb'/> ... # virsh migrate --live vm1 qemu+ssh://target/system error: internal error: Unknown CPU feature sha-ni Could you please help to check this issue ? Thanks in advance for your reply! Sigh, it looks like a never ending punishment for not storing the actual CPU model used to start a domain with host-model CPU in the XML from the moment we added support for host-model. We could possibly try to avoid adding CPU features which were not available in the old libvirt into the guest CPU definition when replacing host-model for a running domain. However, we need to do it in a very smart way to avoid breaking migration to new libvirt. Could you file a separate bug for that and keep this one just for the CPU model name change? (In reply to Jiri Denemark from comment #7) > Sigh, it looks like a never ending punishment for not storing the actual CPU > model used to start a domain with host-model CPU in the XML from the moment > we added support for host-model. We could possibly try to avoid adding CPU > features which were not available in the old libvirt into the guest CPU > definition when replacing host-model for a running domain. However, we need > to do it in a very smart way to avoid breaking migration to new libvirt. > > Could you file a separate bug for that and keep this one just for the CPU > model name change? Okay, I have filed a new bug 1539484 to track this new issue, and thanks for your reply. And verify this bug with comment 6, since the libvirt update the guest cpu to a old qemu supported cpu model, this cpu-model won't broke the migration. 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://access.redhat.com/errata/RHEA-2018:0704 |
Description of problem: libvirt change the running guest cpu model after upgrade, and it will introduce some compatibility issues Version-Release number of selected component (if applicable): libvirt-3.9.0-5.el7.x86_64 How reproducible: 100% Steps to Reproduce: 1. prepare a EPYC host: # lscpu Architecture: x86_64 CPU op-mode(s): 32-bit, 64-bit Byte Order: Little Endian CPU(s): 128 On-line CPU(s) list: 0-127 Thread(s) per core: 2 Core(s) per socket: 32 Socket(s): 2 NUMA node(s): 8 Vendor ID: AuthenticAMD CPU family: 23 Model: 1 Model name: AMD EPYC 7601 32-Core Processor 2. install 7.3 libvirt and qemu: 3. start a guest which use host-model: # virsh dumpxml r7-mig ... <cpu mode='host-model'> <model fallback='allow'/> ... # virsh start r7-mig Domain r7-mig started 4. recheck xml: # virsh dumpxml r7-mig ... <cpu mode='host-model'> <model fallback='allow'/> <numa> <cell id='0' cpus='0-4' memory='524288' unit='KiB'/> <cell id='1' cpus='5-9' memory='524288' unit='KiB'/> </numa> </cpu> # virsh dumpxml r7-mig --migratable ... <cpu mode='host-model' match='exact'> <model fallback='allow'>Opteron_G3</model> <vendor>AMD</vendor> <feature policy='require' name='vme'/> <feature policy='require' name='ht'/> <feature policy='require' name='pclmuldq'/> <feature policy='require' name='ssse3'/> <feature policy='require' name='fma'/> <feature policy='require' name='sse4.1'/> <feature policy='require' name='sse4.2'/> <feature policy='require' name='movbe'/> <feature policy='require' name='aes'/> <feature policy='require' name='xsave'/> <feature policy='require' name='osxsave'/> <feature policy='require' name='avx'/> <feature policy='require' name='f16c'/> <feature policy='require' name='rdrand'/> <feature policy='require' name='arat'/> <feature policy='require' name='fsgsbase'/> <feature policy='require' name='bmi1'/> <feature policy='require' name='avx2'/> <feature policy='require' name='smep'/> <feature policy='require' name='bmi2'/> <feature policy='require' name='rdseed'/> <feature policy='require' name='adx'/> <feature policy='require' name='smap'/> <feature policy='require' name='clflushopt'/> <feature policy='require' name='xsaveopt'/> <feature policy='require' name='xsavec'/> <feature policy='require' name='xgetbv1'/> <feature policy='require' name='mmxext'/> <feature policy='require' name='fxsr_opt'/> <feature policy='require' name='pdpe1gb'/> <feature policy='require' name='cmp_legacy'/> <feature policy='require' name='extapic'/> <feature policy='require' name='cr8legacy'/> <feature policy='require' name='3dnowprefetch'/> <feature policy='require' name='osvw'/> <feature policy='require' name='skinit'/> <feature policy='require' name='wdt'/> <feature policy='require' name='tce'/> <feature policy='require' name='topoext'/> <feature policy='require' name='perfctr_core'/> <feature policy='require' name='perfctr_nb'/> <numa> <cell id='0' cpus='0-4' memory='524288' unit='KiB'/> <cell id='1' cpus='5-9' memory='524288' unit='KiB'/> </numa> </cpu> ... 5. update libvirt and qemu to 7.5: 6. recheck guest xml, libvirt change the model to EPYC: # virsh dumpxml r7-mig ... <cpu mode='custom' match='exact' check='full'> <model fallback='allow'>EPYC</model> <vendor>AMD</vendor> <feature policy='disable' name='ht'/> <feature policy='disable' name='osxsave'/> <feature policy='require' name='cmp_legacy'/> <feature policy='disable' name='extapic'/> <feature policy='disable' name='skinit'/> <feature policy='disable' name='wdt'/> <feature policy='disable' name='tce'/> <feature policy='disable' name='topoext'/> <feature policy='disable' name='perfctr_core'/> <feature policy='disable' name='perfctr_nb'/> <feature policy='disable' name='monitor'/> <feature policy='require' name='x2apic'/> <feature policy='require' name='hypervisor'/> <feature policy='disable' name='sha-ni'/> <feature policy='disable' name='rdtscp'/> <feature policy='disable' name='svm'/> ... # virsh dumpxml r7-mig --migratable ... <cpu mode='custom' match='exact' check='partial'> <model fallback='allow'>EPYC</model> <vendor>AMD</vendor> <feature policy='require' name='ht'/> <feature policy='require' name='osxsave'/> <feature policy='require' name='cmp_legacy'/> <feature policy='require' name='extapic'/> <feature policy='require' name='skinit'/> <feature policy='require' name='wdt'/> <feature policy='require' name='tce'/> <feature policy='require' name='topoext'/> <feature policy='require' name='perfctr_core'/> <feature policy='require' name='perfctr_nb'/> ... 7. this guest cannot migrate to 7.3 since there is no EPYC cpu model: # virsh migrate r7-mig qemu+ssh://target/system --live error: internal error: Unknown CPU model EPYC 8. and after managedsave and restart, libvirt will use EPYC cpu model to build qemu command line: # ps aux|grep qemu ... -cpu Opteron_G3,+vme,+ht,+pclmuldq,+ssse3,+fma,+sse4.1,+sse4.2,+movbe,+aes,+xsave,+osxsave,+avx,+f16c,+rdrand,+arat,+fsgsbase,+bmi1,+avx2,+smep,+bmi2,+rdseed,+adx,+smap,+clflushopt,+xsaveopt,+xsavec,+xgetbv1,+mmxext,+fxsr_opt,+pdpe1gb,+cmp_legacy,+extapic,+cr8legacy,+3dnowprefetch,+osvw,+skinit,+wdt,+tce,+topoext,+perfctr_core,+perfctr_nb # virsh managedsave r7-mig Domain r7-mig state saved by libvirt # virsh start r7-mig Domain r7-mig started # ps aux|grep qemu -cpu EPYC,ht=off,osxsave=off,cmp_legacy=on,extapic=off,skinit=off,wdt=off,tce=off,topoext=off,perfctr_core=off,perfctr_nb=off,monitor=off,x2apic=on,hypervisor=on,sha-ni=off,rdtscp=off,svm=off Actual results: libvirt change the cpu model after upgrade and cause the xml for migrate changed Expected results: Don't change the guest cpu model after upgrade Additional info: From bug 1481252 comment 10: that's unfortunately caused by the older libvirt which did not store the actual CPU configuration it used to start the domain. Thus the libvirt needs to guess it. But apparently we don't do it the way we should. We should ask the running QEMU what CPU model it supports and only use one of the reported models.