Portworx Backup Job Pods

Applicable to both Classic and Federated modes

A Kubernetes Job is a controller that ensures a pod (or multiple pods) completes a finite task successfully. Unlike long-running deployments, Jobs are designed to run once and exit, making them ideal for batch or one-off tasks.

How job pods work

When a Job is created, it spawns one or more pods to complete the specified task. Once the task completes successfully, the Job is marked as complete, and the pods are either kept for logs or auditing or cleaned up depending on the settings.

If the pod fails, Kubernetes can retry the Job according to the spec. For example: backoffLimit, restartPolicy and so on

Types of Kubernetes job patterns

• Single Pod Job: runs once and exits

• Parallel Jobs: run multiple pods simultaneously (for example, data splitting)

• Indexed Jobs (with completions): useful for parallel tasks needing tracking

Portworx Backup cluster job pods

Operation	Job Pod	Purpose
Install	Pre install hook	Prepare environment before installing components
	Post install hook	Perform validation/configuration tasks after install
Upgrade	Pre upgrade hook	Backup or prepare system before applying upgrade
	Post upgrade hook	Verify upgrade, clean up temp data, update configurations
Post install	Post install hook	Finalize setup post install
Cron job	Quick maintenance	Perform lightweight scheduled maintenance tasks. Not supported in Federated mode.
	Full maintenance	Perform comprehensive health checks or data maintenance. Not supported in Federated mode.

Application cluster job pods

Stork is a long-running service deployed on the application cluster and is not a Kubernetes Job pod. In Portworx Backup, Stork orchestrates backup and restore operations by initiating, scheduling, and managing the complete workflow. For more information, see Stork.

The following table lists the Job pods that Portworx Backup creates on the application cluster:

note

In Federated mode, backup and restore operations are fully supported. However, Stork on each application cluster handles the backup and restore workflow directly, so the Job pods listed below are not created. These pods apply to Classic mode only.

Operation	Job Pod	Purpose	Mode
Backup	Backup Pod	Performs backup operations: KDMP backup: backup pod runs on the live PVC and uploads the backup to the backup location. CSI+Offload backup: the snapshot is restored to a PVC, and during the upload phase to the backup location, the pod runs on the restored PVC which is cleaned up after the upload is complete.	Classic mode only
Restore	Restore Pod	Restores KDMP backup present in the backup location.	Classic mode only
Command Execution	Cmd Executor pod	Executes pre and post exec rules during backup	Classic and Federated modes
NFS Operations	NFS Executor Pod	Resource upload/download when backup/restore is done to an NFS backup location.	Classic mode only