Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

handle all restart policies #1649

Merged
merged 4 commits into from
Dec 30, 2022
Merged

handle all restart policies #1649

Changes from all commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
8f7f950
handle all restart policies
abin-thomas-by Aug 10, 2022
c6c585e
fix lint error
abin-thomas-by Dec 29, 2022
ebeb28f
Merge branch 'kubeflow:master' into pod_policy
abin-thomas-by Dec 29, 2022
5cc8691
gofmt whitespace diff
abinthomasonline Dec 30, 2022
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
35 changes: 18 additions & 17 deletions pkg/controller.v1/tensorflow/tfjob_controller.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -784,25 +784,26 @@ func (r *TFJobReconciler) ReconcilePods(
}
}
// Check if the pod is retryable.
if spec.RestartPolicy == commonv1.RestartPolicyExitCode {
if pod.Status.Phase == v1.PodFailed && train_util.IsRetryableExitCode(exitCode) {
logger.Infof("Need to restart the pod: %v.%v", pod.Namespace, pod.Name)
if err := r.PodControl.DeletePod(pod.Namespace, pod.Name, tfJob); err != nil {
return err
}
if pod.Status.Phase == v1.PodFailed &&
(spec.RestartPolicy == commonv1.RestartPolicyExitCode && train_util.IsRetryableExitCode(exitCode) ||
spec.RestartPolicy == commonv1.RestartPolicyOnFailure ||
spec.RestartPolicy == commonv1.RestartPolicyAlways) {
logger.Infof("Need to restart the pod: %v.%v", pod.Namespace, pod.Name)
if err := r.PodControl.DeletePod(pod.Namespace, pod.Name, tfJob); err != nil {
return err
}

// with common library framework, we have to handle restart status here
// or we won't know which replica has been restarted in updateJobStatus after reconciling all replicas
msg := fmt.Sprintf("TFJob %s is restarting because %s replica(s) failed.",
tfJob.Name, rtype)
r.Recorder.Event(tfJob, corev1.EventTypeWarning, tfJobRestartingReason, msg)
err := commonutil.UpdateJobConditions(jobStatus, commonv1.JobRestarting, tfJobRestartingReason, msg)
if err != nil {
commonutil.LoggerForJob(tfJob).Infof("Append tfjob condition error: %v", err)
return err
}
trainingoperatorcommon.RestartedJobsCounterInc(tfJob.Namespace, kubeflowv1.TFJobFrameworkName)
// with common library framework, we have to handle restart status here
// or we won't know which replica has been restarted in updateJobStatus after reconciling all replicas
msg := fmt.Sprintf("TFJob %s is restarting because %s replica(s) failed.",
tfJob.Name, rtype)
r.Recorder.Event(tfJob, corev1.EventTypeWarning, tfJobRestartingReason, msg)
err := commonutil.UpdateJobConditions(jobStatus, commonv1.JobRestarting, tfJobRestartingReason, msg)
if err != nil {
commonutil.LoggerForJob(tfJob).Infof("Append tfjob condition error: %v", err)
return err
}
trainingoperatorcommon.RestartedJobsCounterInc(tfJob.Namespace, kubeflowv1.TFJobFrameworkName)
}

updateJobReplicaStatuses(jobStatus, rtype, pod)
Expand Down