Red Hat Bugzilla – Bug 1235180
guest will have broken settings if we cold-unplug a vcpu which included in some domain vcpu sched
Last modified: 2016-11-03 14:19:11 EDT
Description of problem: guest will have broken settings if we delete a vcpu which included in some domain vcpu sched Version-Release number of selected component (if applicable): libvirt-1.2.16-1.el7.x86_64 How reproducible: 100% Steps to Reproduce: 1. prepare a guest like this: <vcpu placement='static' cpuset='0-1'>5</vcpu> ... <cputune> <emulatorpin cpuset='1-3'/> <vcpusched vcpus='0-2,4' scheduler='idle'/> <iothreadsched iothreads='1-2' scheduler='idle'/> </cputune> 2. use setvcpus remove vcpu 4: # virsh setvcpus test4 --maximum 4 --config 3. recheck xml <vcpu placement='static' cpuset='0-1'>4</vcpu> ... <cputune> <emulatorpin cpuset='1-3'/> <vcpusched vcpus='0-2,4' scheduler='idle'/> <iothreadsched iothreads='1-2' scheduler='idle'/> </cputune> 4. restart libvirtd and check guest # service libvirtd restart Redirecting to /bin/systemctl restart libvirtd.service # virsh dumpxml test4 error: failed to get domain 'test4' error: Domain not found: no domain with matching name 'test4' Actual results: guest will have broken settings if we delete a vcpu which included in some domain vcpu sched Expected results: remove the use of vcpu 4 in vcpusched in step 3 Additional info:
I'm working on a refactor of data structures that hold info for vCPUs so this bug should be fixed along with that series.
Fixed upstream: commit 99c5fe0e7c26c08103415248ffef1f5acb81ddc7 Author: Peter Krempa <pkrempa@redhat.com> Date: Tue Jan 12 13:12:05 2016 +0100 conf: Don't store vcpusched orthogonally to other vcpu info Due to bad design the vcpu sched element is orthogonal to the way how the data belongs to the corresponding objects. Now that vcpus are a struct that allow to store other info too, let's convert the data to the sane structure. The helpers for the conversion are made universal so that they can be reused for iothreads too. This patch also resolves https://bugzilla.redhat.com/show_bug.cgi?id=1235180 since with the correct storage approach you can't have dangling data. v1.3.1-159-g99c5fe0
Verify this bug on libvirt-2.0.0-5.el7.x86_64: 1. prepare a guest xml like this: # virsh dumpxml r7 ... <vcpu placement='static'>20</vcpu> <iothreads>1</iothreads> <cputune> <vcpusched vcpus='0-2,18' scheduler='idle'/> <iothreadsched iothreads='1' scheduler='idle'/> </cputune> ... <cpu mode='host-model'> <model fallback='allow'/> <numa> <cell id='0' cpus='0-4' memory='524288' unit='KiB' memAccess='shared'/> <cell id='1' cpus='5-8' memory='524288' unit='KiB' memAccess='shared'/> </numa> </cpu> 2. use setvcpus remove some vcpus: # virsh setvcpus r7 10 --config --maximum 3. recheck xml: # virsh dumpxml r7 ... <vcpu placement='static'>10</vcpu> <iothreads>1</iothreads> <cputune> <vcpusched vcpus='0-2' scheduler='idle'/> <iothreadsched iothreads='1' scheduler='idle'/> </cputune> ... 4. restart libvirtd and recheck guest xml: # service libvirtd restart Redirecting to /bin/systemctl restart libvirtd.service # virsh dumpxml r7 ... <vcpu placement='static'>10</vcpu> <iothreads>1</iothreads> <cputune> <vcpusched vcpus='0-2' scheduler='idle'/> <iothreadsched iothreads='1' scheduler='idle'/> </cputune> ...
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://rhn.redhat.com/errata/RHSA-2016-2577.html