Federation V1, the current Kubernetes federation API which reuses the Kubernetes API resources ‘as is’, is currently considered alpha for many of its features. There is no clear path to evolve the API to GA; however, there is a
Federation V2effort in progress to implement a dedicated federation API apart from the Kubernetes API. The details are available at sig-multicluster community page.
This guide explains how to use jobs in the federation control plane.
Jobs in the federation control plane (referred to as “federated jobs” in this guide) are similar to the traditional Kubernetes jobs, and provide the same functionality. Creating jobs in the federation control plane ensures that the desired number of parallelism and completions exist across the registered clusters.
This guide assumes that you have a running Kubernetes Cluster Federation installation. If not, then head over to the federation admin guide to learn how to bring up a cluster federation (or have your cluster administrator do this for you). Other tutorials, such as Kelsey Hightower’s Federated Kubernetes Tutorial, might also help you create a Federated Kubernetes cluster.
The API for federated jobs is fully compatible with the API for traditional Kubernetes jobs. You can create a job by sending a request to the federation apiserver.
You can do that using kubectl by running:
kubectl --context=federation-cluster create -f myjob.yaml
The ‘–context=federation-cluster’ flag tells kubectl to submit the request to the federation API server instead of sending it to a Kubernetes cluster.
Once a federated job is created, the federation control plane creates a job in all underlying Kubernetes clusters. You can verify this by checking each of the underlying clusters, for example:
kubectl --context=gce-asia-east1a get job myjob
The previous example assumes that you have a context named
configured in your client for your cluster in that zone.
The jobs in the underlying clusters match the federated job except in the number of parallelism and completions. The federation control plane ensures that the sum of the parallelism and completions in each cluster matches the desired number of parallelism and completions in the federated job.
By default, parallelism and completions are spread equally in all underlying clusters. For example:
if you have 3 registered clusters and you create a federated job with
spec.parallelism = 9 and
spec.completions = 18, then each job in the 3 clusters has
spec.parallelism = 3 and
spec.completions = 6.
To modify the number of parallelism and completions in each cluster, you can specify
as an annotation with key
on the federated job.
You can update a federated job as you would update a Kubernetes job; however, for a federated job, you must send the request to the federation API server instead of sending it to a specific Kubernetes cluster. The federation control plane ensures that whenever the federated job is updated, it updates the corresponding job in all underlying clusters to match it.
If your update includes a change in number of parallelism and completions, the federation control plane changes the number of parallelism and completions in underlying clusters to ensure that their sum remains equal to the number of desired parallelism and completions in federated job.
You can delete a federated job as you would delete a Kubernetes job; however, for a federated job, you must send the request to the federation API server instead of sending it to a specific Kubernetes cluster.
For example, with kubectl:
kubectl --context=federation-cluster delete job myjob
Note: Deleting a federated job will not delete the corresponding jobs from underlying clusters. You must delete the underlying jobs manually.