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

Handle all ray job statuses #4389

Merged
merged 2 commits into from
Nov 10, 2023
Merged

Handle all ray job statuses #4389

Changes from 1 commit
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
5268391
Handle all ray job statuses
EngHabu Nov 9, 2023
8a3ca32
PR Comments
EngHabu Nov 10, 2023
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
16 changes: 11 additions & 5 deletions flyteplugins/go/tasks/plugins/k8s/ray/ray.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -470,7 +470,7 @@ func (plugin rayJobResourceHandler) GetTaskPhase(ctx context.Context, pluginCont
return pluginsCore.PhaseInfoQueued(time.Now(), pluginsCore.DefaultPhaseVersion, "Scheduling"), nil
}

// Kuberay creates a Ray cluster first, and then submits a Ray job to the cluster
// KubeRay creates a Ray cluster first, and then submits a Ray job to the cluster
switch rayJob.Status.JobDeploymentStatus {
case rayv1alpha1.JobDeploymentStatusInitializing:
return pluginsCore.PhaseInfoInitializing(rayJob.CreationTimestamp.Time, pluginsCore.DefaultPhaseVersion, "cluster is creating", info), nil
Expand All @@ -480,7 +480,7 @@ func (plugin rayJobResourceHandler) GetTaskPhase(ctx context.Context, pluginCont
case rayv1alpha1.JobDeploymentStatusFailedJobDeploy:
reason := fmt.Sprintf("Failed to submit Ray job %s with error: %s", rayJob.Name, rayJob.Status.Message)
return pluginsCore.PhaseInfoFailure(flyteerr.TaskFailedWithError, reason, info), nil
case rayv1alpha1.JobDeploymentStatusWaitForDashboard:
case rayv1alpha1.JobDeploymentStatusWaitForDashboard, rayv1alpha1.JobDeploymentStatusFailedToGetJobStatus:
return pluginsCore.PhaseInfoRunning(pluginsCore.DefaultPhaseVersion, info), nil
case rayv1alpha1.JobDeploymentStatusRunning, rayv1alpha1.JobDeploymentStatusComplete:
switch rayJob.Status.JobStatus {
Expand All @@ -489,12 +489,18 @@ func (plugin rayJobResourceHandler) GetTaskPhase(ctx context.Context, pluginCont
return pluginsCore.PhaseInfoFailure(flyteerr.TaskFailedWithError, reason, info), nil
case rayv1alpha1.JobStatusSucceeded:
return pluginsCore.PhaseInfoSuccess(info), nil
case rayv1alpha1.JobStatusPending, rayv1alpha1.JobStatusRunning:
case rayv1alpha1.JobStatusPending, rayv1alpha1.JobStatusRunning, rayv1alpha1.JobStatusStopped:
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Stopped sounds weird to be running

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

It's not currently being used in the operator... so I took a guess it's a temporary state (Succeeded/Failed are terminal)... hence why I marked it as Running. That said, I think it's ok for now to add it to the default handler of failing as an unknown/not handled state..

return pluginsCore.PhaseInfoRunning(pluginsCore.DefaultPhaseVersion, info), nil
default:
// We already handle all known job status, so this should never happen unless a future version of ray
// introduced a new job status.
return pluginsCore.PhaseInfoUndefined, fmt.Errorf("unknown job status: %s", rayJob.Status.JobStatus)
}
default:
// We already handle all known deployment status, so this should never happen unless a future version of ray
// introduced a new job status.
return pluginsCore.PhaseInfoUndefined, fmt.Errorf("unknown job deployment status: %s", rayJob.Status.JobDeploymentStatus)
}

return pluginsCore.PhaseInfoUndefined, nil
}

func init() {
Expand Down
Loading