consider making pipeline runtime image image_name reference sha256 digest format

[shalberd]

related to

opendatahub-io-contrib/workbench-images#44

docker image tag can be put in tags array of json.
the image_name should be in digest format

for all runtime image jsons at notebooks/jupyter/datascience/

so as to be compatible with restricted network environments and ImageContentSourcePolicy. Important for restricted network environments / airgapped. Yes, users can add additional runtimes via GUI per user notebook / PVC, but the default runtimes as delived via ODH in the docker jupyter datascience images should be in the correct format as well.

image_name is used for pulling images, tags are shown in Elyra GUI, just like display_name.

i.e.

{
    "display_name": "CUDA Datascience with Python 3.11 (CentOS Stream 9)",
    "metadata": {
        "tags": ["cuda-runtime-datascience-c9s-py39_RELEASE_latest"],
        "display_name": "CUDA Datascience with Python 3.11 (CentOS Stream 9)",
        "image_name": "quay.io/opendatahub-contrib/workbench-images@sha256:bd565c0e8b4e71d7ae3fd556a2f874e6251d80db1c4c36bfe1d490f04df2bdb1"
    },
    "schema_name": "runtime-image"
}

[harshad16]

Acceptance criteria

• Change the from tag to sha256 format. ex: link

[shalberd]

@harshad16 yes, and for readability in Elyra GUI, you can put the tag that was originally part of the docker image image_name in the tags array instead, as an element of that. https://github.com/opendatahub-io/notebooks/blob/main/jupyter/datascience/ubi9-python-3.9/runtime-images/datascience-ubi9-py39.json#L4

I wonder if image pull policy is also a part of that json. If yes, since we would now be at unique hashes, we could use IfNotPresent, saving a lot of time during pipeline task container startup potentially. i.e. add pull_policy to json in metadata

https://github.com/elyra-ai/elyra/blob/main/docs/source/user_guide/runtime-image-conf.md?plain=1#L213

"pull_policy": "IfNotPresent"

full example:

{
    "display_name": "CUDA Datascience with Python 3.11 (CentOS Stream 9)",
    "metadata": {
        "tags": ["cuda-runtime-datascience-c9s-py39_2023c_latest"],
        "display_name": "CUDA Datascience with Python 3.11 (CentOS Stream 9)",
        "image_name": "quay.io/opendatahub-contrib/workbench-images@sha256:bd565c0e8b4e71d7ae3fd556a2f874e6251d80db1c4c36bfe1d490f04df2bdb1",
        "pull_policy": "IfNotPresent"
    },
    "schema_name": "runtime-image"
}

[harshad16]

This feature is completed with the PR: #286
Marking this as complete, please check the next opendatahub release to see this in effect.
Closing the issue, thanks all for the work 💯