github
/
k3s
mirror of https://github.com/k3s-io/k3s
1
0
Fork 0
Code Issues Projects Releases Wiki Activity

Merge pull request #43030 from yujuhong/rm_corrupted_checkpoint 

Automatic merge from submit-queue (batch tested with PRs 42747, 43030)

dockershim: remove corrupted sandbox checkpoints

This is a workaround to ensure that kubelet doesn't block forever when
the checkpoint is corrupted.

This is a workaround for #43021
pull/6/head
Kubernetes Submit Queue 2017-03-14 22:56:20 -07:00 committed by GitHub
parent 1b4433d4c3 035afab901
commit 222f69cf3c
1 changed files with 14 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

16
pkg/kubelet/dockershim/docker_sandbox.go
View File

@ -144,6 +144,16 @@ func (ds *dockerService) StopPodSandbox(podSandboxID string) error {
glog.Warningf("Both sandbox container and checkpoint for id %q could not be found. "+
"Proceed without further sandbox information.", podSandboxID)
} else {
if checkpointErr == errors.CorruptCheckpointError {
// Remove the corrupted checkpoint so that the next
// StopPodSandbox call can proceed. This may indicate that
// some resources won't be reclaimed.
// TODO (#43021): Fix this properly.
glog.Warningf("Removing corrupted checkpoint %q: %+v", podSandboxID, *checkpoint)
if err := ds.checkpointHandler.RemoveCheckpoint(podSandboxID); err != nil {
glog.Warningf("Unable to remove corrupted checkpoint %q: %v", podSandboxID, err)
}
}
return utilerrors.NewAggregate([]error{
fmt.Errorf("failed to get checkpoint for sandbox %q: %v", podSandboxID, checkpointErr),
fmt.Errorf("failed to get sandbox status: %v", statusErr)})
@ -393,8 +403,10 @@ func (ds *dockerService) ListPodSandbox(filter *runtimeapi.PodSandboxFilter) ([]
glog.Errorf("Failed to retrieve checkpoint for sandbox %q: %v", id, err)
if err == errors.CorruptCheckpointError {
glog.V(2).Info("Removing corrupted checkpoint %q: %+v", id, *checkpoint)
ds.checkpointHandler.RemoveCheckpoint(id)
glog.Warningf("Removing corrupted checkpoint %q: %+v", id, *checkpoint)
if err := ds.checkpointHandler.RemoveCheckpoint(id); err != nil {
glog.Warningf("Unable to remove corrupted checkpoint %q: %v", id, err)
}
}
continue
}