training-operator Add Helm Charts for Kubeflow Trainer V2

Dec 11 '20 10:12 gaocegege

I would like to work on this issue. Since I'm new here, can someone guide me?

Dec 11 '20 14:12 sanjeepan23

Thanks, @goat023

You can have a look at other operator's helm chart, like tidb-operator or etcd-operator. It should be similar to them.

Dec 12 '20 07:12 gaocegege

There's also helm chart for mpi-operator for reference but probably needs some update as well: https://github.com/kubeflow/mpi-operator/tree/master/hack/helm/mpi-operator

Dec 12 '20 20:12 terrytangyuan

Thanks, I'll try

Dec 13 '20 03:12 sanjeepan23

where can I get the maintainer's name and email address for this repo

Dec 13 '20 04:12 sanjeepan23

https://github.com/kubeflow/tf-operator/blob/master/OWNERS

Dec 14 '20 02:12 gaocegege

@goat023 Hi, is there any progress?

Jan 18 '21 06:01 gaocegege

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.

Jun 02 '21 17:06 stale[bot]

Do you still need one?

Jun 15 '22 12:06 ehudyonasi

Inorder to ensure that helm charts are in sync, we can add one extra check in Github Actions.

Jul 18 '22 05:07 johnugeorge

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.

Oct 04 '23 20:10 github-actions[bot]

This issue has been automatically closed because it has not had recent activity. Please comment "/reopen" to reopen it.

Oct 24 '23 20:10 github-actions[bot]

More and more Kubernetes projects are moved toward Helm charts for the deployment:

Kueue: https://github.com/kubernetes-sigs/kueue/tree/main/charts/kueue
JobSet: https://github.com/kubernetes-sigs/jobset/issues/726

I think, we should consider to create Helm Chart for at least Kubeflow Training V2 WDYT @kubeflow/wg-training-leads @ChenYi015 ?

Dec 19 '24 22:12 andreyvelich

/remove-lifecycle stale

Dec 19 '24 22:12 andreyvelich

I strongly agree with adding a Helm chart for the training operator, and I am willing to maintain it, drawing on my experience with the Spark operator Helm chart.

Dec 20 '24 14:12 ChenYi015

Great, thank you @ChenYi015! I would also suggest to take a look at the sync script that @tenzen-y uses for Kueue project: https://github.com/kubeflow/training-operator/pull/2263#issuecomment-2407994501

Dec 20 '24 14:12 andreyvelich

+1 as long as we can sustain the maintenance

Dec 22 '24 16:12 terrytangyuan

Let's include this feature as part of Release 2.0, so users deploy Kubeflow Training V2 with Helm Charts. @ChenYi015 Do you have bandwidth to work on this and review @tenzen-y script to keep Helm Charts and Kustomize manifests in sync: https://github.com/kubernetes-sigs/kueue/tree/main/hack ?

Jan 08 '25 18:01 andreyvelich

Do you have bandwidth to work on this and review @tenzen-y script to keep Helm Charts and Kustomize manifests in sync: https://github.com/kubernetes-sigs/kueue/tree/main/hack ?

I will do it when I have time.

Jan 09 '25 11:01 ChenYi015

Hi @ChenYi015, it would be nice if we could prepare Helm Charts for Kubeflow Trainer before we release the first version. Do you think you still have time to help us or we should ask community for the help ?

Feb 11 '25 20:02 andreyvelich

@andreyvelich Ok, I will update the helm charts ASAP. When are we going to release the first version of V2?

Feb 12 '25 02:02 ChenYi015

@ChenYi015 We are planning to release V2 right after we implement the MPI Plugin: https://github.com/kubeflow/trainer/pull/2394

Feb 12 '25 15:02 andreyvelich

/assign @ChenYi015 cc @andreyvelich

Feb 14 '25 03:02 varodrig