Just to add, Once this happens its breaks standard libvirt tooling:
virsh cpu-stats 7 error: Failed to retrieve CPU statistics for domain 'instance-00000681' error: Unable to read from '/sys/fs/cgroup/cpu,cpuacct/machine.slice/machine-qemu\x2dinstance\x2d00000681.scope/cpuacct.usage_percpu': No such file or directory Apparent steps to reproduce: systemctl daemon-restart systemctl restart (some unrelated service) Within 60 seconds /sys/fs/cgroup/cpu,cpuacct/machine.slice folder is purged. ____________________________________________ Kris Lindgren Senior Linux Systems Engineer GoDaddy, LLC. From: Michael Dorman <[email protected]<mailto:[email protected]>> Date: Tuesday, December 2, 2014 at 11:44 AM To: OpenStack Operators <[email protected]<mailto:[email protected]>> Subject: [Openstack-operators] libvirt cgroups getting removed under Cent 7/systemd We recently discovered a problem where the cgroups set up by libvirt for VMs are being removed by systemd. It's this bug: https://bugzilla.redhat.com/show_bug.cgi?id=1139223 Curious if anyone else has run into this, and/or has figured out a workaround until it's fixed in upstream CentOS/RHEL. Thanks, Mike
_______________________________________________ OpenStack-operators mailing list [email protected] http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-operators
