training-operator icon indicating copy to clipboard operation
training-operator copied to clipboard

Consider container image rename of `kubeflow/storage-initializer`

Open tarilabs opened this issue 1 year ago • 11 comments

What you would like to be added?

This image: https://hub.docker.com/r/kubeflow/storage-initializer from: https://github.com/kubeflow/training-operator/blob/master/sdk/python/kubeflow/storage_initializer/Dockerfile

is generically called: kubeflow/storage-initializer to integrate with KServe as a Custom Storage Initializer.

Kindly consider renaming it something more specific to avoid end-user confusion?

Why is this needed?

From Model Registry, we are also now producing a KServe CSI image: https://hub.docker.com/r/kubeflow/model-registry-storage-initializer named specifically kubeflow/model-registry-storage-initializer to serve integration needs of Model Registry <-> KServe .

For more details, see this guide.

I believe making the image name specific to <capability>-storage-initializer will avoid confusion in the long term for Kubeflow end-users, also if other KF components will end up providing their own requirements of CSI !

wdyt @andreyvelich @terrytangyuan ?

Love this feature?

Give it a 👍 We prioritize the features with most 👍

tarilabs avatar Jul 25 '24 13:07 tarilabs

Thanks for creating this @tarilabs! Yes, we might want to find better name (e.g. training-storage-initializer) cc @johnugeorge @deepanker13

andreyvelich avatar Jul 30 '24 17:07 andreyvelich

Yes agreed that we should have more unique names based on capabilities.

terrytangyuan avatar Jul 31 '24 01:07 terrytangyuan

shall we rename it to kubeflow/model-dataset-downloader ? cc @andreyvelich @terrytangyuan

deepanker13 avatar Aug 01 '24 23:08 deepanker13

shall we rename it to kubeflow/model-dataset-downloader ? cc @andreyvelich @terrytangyuan

I think that training- prefix would be better so that we can avoid conflicting with other projects.

tenzen-y avatar Aug 02 '24 05:08 tenzen-y

Yes, we might want to add training-. Additionally, after V2, we are planning to separate dataset-initializer and model-initializer in 2 different containers: https://github.com/kubeflow/training-operator/pull/2171

andreyvelich avatar Aug 02 '24 11:08 andreyvelich

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.

github-actions[bot] avatar Oct 31 '24 15:10 github-actions[bot]

still valid imho.

tarilabs avatar Oct 31 '24 15:10 tarilabs

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.

github-actions[bot] avatar Jan 29 '25 20:01 github-actions[bot]

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

github-actions[bot] avatar Feb 18 '25 20:02 github-actions[bot]

/reopen

tarilabs avatar Feb 19 '25 13:02 tarilabs

@tarilabs: Reopened this issue.

In response to this:

/reopen

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository.

google-oss-prow[bot] avatar Feb 19 '25 13:02 google-oss-prow[bot]

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.

github-actions[bot] avatar May 20 '25 20:05 github-actions[bot]

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

github-actions[bot] avatar Jun 10 '25 00:06 github-actions[bot]

Considering the images now live https://github.com/kubeflow/trainer/pkgs/container/training-v1%2Fstorage-initializer

it can be left closed.

tarilabs avatar Jun 10 '25 06:06 tarilabs