[BUG]: `pipeline requires workspace with name "model-workspace" be provided by pipelinerun`

[adelton]

Details

Describe the bug

Following the pipelines/README.md, the build-container-image-tensorflow-housing-t2mtr PipelineRun fails with

PipelineRun model-to-edge/build-container-image-tensorflow-housing-t2mtr doesn't bind Pipeline model-to-edge/build-container-image's Workspaces correctly: pipeline requires workspace with name "model-workspace" be provided by pipelinerun

To Reproduce

1. Get a fresh OCP cluster.
2. As admin, install the Red Hat OpenShift Pipelines 1.13.0 operator via a web console.
3. Per pipelines/README.md, as non-admin user, oc new-project model-to-edge.
4. Per pipelines/README.md, edit credentials to the S3:

   cp tekton/build-container-image-pipeline/aws-env.yaml tekton/build-container-image-pipeline/aws-env-real.yaml
   vi tekton/build-container-image-pipeline/aws-env-real.yaml

5. Per pipelines/README.md, as non-admin user, oc create -f tekton/build-container-image-pipeline/aws-env-real.yaml.
6. Per pipelines/README.md, update aws-bucket-name value in both pipelines/tekton/build-container-image-pipeline/build-container-image-pipelinerun-bike-rentals.yaml and pipelines/tekton/build-container-image-pipeline/build-container-image-pipelinerun-tensorflow-housing.yaml to point to an S3 bucket with copies of the models.
7. Per pipelines/README.md, as non-admin user,

   oc apply -k tekton/build-container-image-pipeline/
   oc create -f tekton/build-container-image-pipeline/build-container-image-pipelinerun-bike-rentals.yaml
   # and / or
   oc create -f tekton/build-container-image-pipeline/build-container-image-pipelinerun-tensorflow-housing.yaml

8. As admin on the web console, check the Pipelines and PipelineRuns in the model-to-edge namespace.

See the build-container-image-bike-rentals-9vw2d PipelineRun eventually reported as Succeeded.

See the build-container-image-tensorflow-housing-t2mtr reported as Failed, with a log snippet message

PipelineRun model-to-edge/build-container-image-tensorflow-housing-t2mtr doesn't bind Pipeline model-to-edge/build-container-image's Workspaces correctly: pipeline requires workspace with name "model-workspace" be provided by pipelinerun

Expected behavior

Both PipelineRuns (build-container-image-bike-rentals-9vw2d and build-container-image-tensorflow-housing-t2mtr) end up in the Succeeded state.

Screenshots (if applicable)

[adelton]

@jackdelahunt, @LaVLaS, since the model-workspace was introduced via https://github.com/opendatahub-io/ai-edge/pull/115, it seems that PR might be the source of the regression. When I reset my repo to 63ef2dcbd3a0e82afe81f0606abf53364ba7a99b (before https://github.com/opendatahub-io/ai-edge/pull/115 got merged), both the PipelineRuns work as expected (unless one hits https://github.com/opendatahub-io/ai-edge/issues/175 as well).

[adelton]

Hello @MarianMacik, was this bug addressed by the 77be5caa034379546c265eaa19a542101f789ea3 commit of https://github.com/opendatahub-io/ai-edge/pull/191, and thus we could mark this issue as fixed?

[MarianMacik]

@adelton Sorry, I have just noticed it now. Yes, this bug was addressed. Fixed by #191 in https://github.com/opendatahub-io/ai-edge/commit/77be5caa034379546c265eaa19a542101f789ea3.