vbatts/cri-o
1
0
Fork 0
Code Issues Pull requests Releases Wiki Activity

oci: Remove useless crio-conmon- cgroup deletion

Browse source 
It always fails because conmon is still there.
But more importantly it adds a 2 seconds delay to the container
creation as we're trying to delete a cgroup but we can't.

With this patch a container creation is down to typically less than
150ms instead of 2+ seconds.

Signed-off-by: Samuel Ortiz <sameo@linux.intel.com>
This commit is contained in:
Samuel Ortiz 2017-10-13 11:27:56 +02:00
parent 436194290a
commit 29121c8c0c
1 changed files with 6 additions and 5 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

11
oci/oci.go
View file

@ -224,11 +224,12 @@ func (r *Runtime) CreateContainer(c *Container, cgroupParent string) (err error)
if err != nil {
logrus.Warnf("Failed to add conmon to cgroupfs sandbox cgroup: %v", err)
} else {
// XXX: this defer does nothing as the cgroup can't be deleted cause
// it contains the conmon pid in tasks
// we need to remove this defer and delete the cgroup once conmon exits
// maybe need a conmon monitor?
defer control.Delete()
// Here we should defer a crio-connmon- cgroup hierarchy deletion, but it will
// always fail as conmon's pid is still there.
// Fortunately, kubelet takes care of deleting this for us, so the leak will
// only happens in corner case where one does a manual deletion of the container
// through e.g. runc. This should be handled by implementing a conmon monitoring
// routine that does the cgroup cleanup once conmon is terminated.
if err := control.Add(cgroups.Process{Pid: cmd.Process.Pid}); err != nil {
logrus.Warnf("Failed to add conmon to cgroupfs sandbox cgroup: %v", err)
}