Bug 1532980
Summary: | guest shutoff after upgrade from 7.3 to 7.5 on skylake host | ||
---|---|---|---|
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: | high | ||
Version: | 7.5 | CC: | dyuan, lhuang, mtessun, rbalakri, salmy, xuzhang, yalzhang |
Target Milestone: | rc | ||
Target Release: | --- | ||
Hardware: | x86_64 | ||
OS: | Linux | ||
Whiteboard: | |||
Fixed In Version: | libvirt-3.9.0-8.el7 | Doc Type: | If docs needed, set a value |
Doc Text: | Story Points: | --- | |
Clone Of: | Environment: | ||
Last Closed: | 2018-04-10 11:04:21 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: |
Description
Luyao Huang
2018-01-10 07:20:19 UTC
The problem is in the fallback attribute: <cpu mode='host-model' check='partial'> <model fallback='forbid'/> </cpu> when libvirtd reconnects to a domain with such CPU configuration started by old libvirt, the host-model CPU gets replaced with a custom one (the one found in host capabilities), but the fallback attribute remains 'forbid'. If libvirtd detects the host CPU as a model which is too new to be supported by the already running QEMU, we need to translate it to a supported model. But this is impossible if fallback == 'forbid'. So the translation fails, reconnecting to the domains fails, and the domain gets killed by libvirtd. We need to reset the fallback attribute to 'allow' when replacing host-model with the CPU model from host capabilities as such CPU is not (by design) tailored to any specific QEMU binary and it is expected to be translated when being passed to QEMU. This is now fixed upstream by commit 237f045d9a9d84a52e645b490d14ad0b66e5b8ed Refs: v3.10.0-182-g237f045d9a Author: Jiri Denemark <jdenemar> AuthorDate: Wed Jan 10 15:56:21 2018 +0100 Commit: Jiri Denemark <jdenemar> CommitDate: Fri Jan 12 10:45:31 2018 +0100 qemu: Ignore fallback CPU attribute on reconnect When reconnecting to a running domain with host-model CPU started by old libvirt which did not store the actual CPU in the status XML, we need to ignore the fallback attribute to make sure we can translate the detected host CPU model to a model which is supported by the running QEMU. https://bugzilla.redhat.com/show_bug.cgi?id=1532980 Signed-off-by: Jiri Denemark <jdenemar> Reviewed-by: Pavel Hrdina <phrdina> Verify this bug on libvirt-3.9.0-8.el7.x86_64: 1. guest xml on 7.3: <cpu mode='host-model'> <model fallback='forbid'/> <numa> <cell id='0' cpus='0-1' memory='512000' unit='KiB'/> <cell id='1' cpus='2-3' memory='512000' unit='KiB'/> </numa> </cpu> 2. start guest: # virsh start vm1 Domain vm1 started 3. update libvirt and qemu to 7.5: 4. check guest status: # virsh list Id Name State ---------------------------------------------------- 1 vm1 running 5. check guest xml # virsh dumpxml vm1 <cpu mode='custom' match='exact' check='full'> <model fallback='allow'>Skylake-Client-IBRS</model> <vendor>Intel</vendor> 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 |