Failed to rerun PipelineRun requires WorkspaceBinding

[chechiachang]

I have a pipeline require a workspace. The pipeline can be triggered by

web hook -> TriggerTemplate (providing VolumeClaimTemplate) -> PipelineRun

The PipelineRun works perfect, but it can't rerun from dashboard. This makes sense since the rerun don't provide required WorkspaceBinding. Is there a better way to provide WorkspaceBinding?

Expected Behavior

Successfully rerun from dashboard (or /rerun api)

Actual Behavior

PipelineRun failed to create TaskRun.

kubectl get pipelinerun
NAME                       SUCCEEDED   REASON            STARTTIME   COMPLETIONTIME
tekton-cd-wn4bw            True        Completed         155m        154m
tekton-cd-wn4bw-r-hbrzf    Unknown                       38m

tkn pipelinerun logs tekton-cd-wn4bw-r-hbrzf
Pipeline still running ...
No logs found

Controller log error:

kubectl logs tekton-pipelines-controller
..."msg":"Reconcile error: error creating TaskRun called tekton-cd-wn4bw-r-hbrzf-git-
clone-q4pbw for PipelineTask generic-git-clone from PipelineRun tekton-cd-wn4bw-r-
hbrzf: admission webhook \"validation.webhook.pipeline.tekton.dev\" denied the request:
validation failed: expected exactly one, got neither: workspace.configmap,
workspace.emptydir, workspace.persistentvolumeclaim, workspace.secret,
workspace.volumeclaimtemplate"...

Steps to Reproduce the Problem

Pipeline

apiVersion: tekton.dev/v1beta1
kind: Pipeline
metadata:
  name: tekton-cd
spec:
  workspaces:
    - name: shared-workspace
  params:
  tasks:
---
apiVersion: triggers.tekton.dev/v1alpha1
kind: TriggerTemplate
metadata:
  name: tekton-cd
  namespace: tekton-pipelines
spec:
  params:
  resourcetemplates:
    - apiVersion: tekton.dev/v1beta1
      kind: PipelineRun
      metadata:
        generateName: tekton-cd-
      spec:
        pipelineRef:
          name: tekton-cd
        workspaces:
          - name: shared-workspace
            volumeClaimTemplate:
              spec:
                accessModes:
                  - ReadWriteOnce
                resources:
                  requests:
                    storage: 1Gi

1. Trigger PipelineRun with webhook & TriggerTemplate -> Succeeded.
2. Rerun the same PipelineRun from dashboard -> Failed.

Additional Info

What I've tried:

1. Use Git PipelineSource: In my case, the workspace contains a git source. The same Pipeline works if I replace workspace with Git PipelineResource.
2. Allocate PVC with kubectl and then use existing pvc as workspace. This block Pipeline from parallel run

workspaces:
- name: myworkspace
  persistentVolumeClaim:
    claimName: my-existing-pvc

• Kubernetes version:

Client Version: version.Info{
	Major: "1",
	Minor: "17",
	GitVersion: "v1.17.3",
	GitCommit: "06ad960bfd03b39c8310aaf92d1e7c12ce618213",
	GitTreeState: "clean",
	BuildDate: "2020-02-11T18:14:22Z",
	GoVersion: "go1.13.6",
	Compiler: "gc",
	Platform: "darwin/amd64"
}
Server Version: version.Info{
	Major: "1",
	Minor: "16+",
	GitVersion: "v1.16.8-gke.15",
	GitCommit: "9cabee15e0922c3b36724de4866a98f6c2da5e6a",
	GitTreeState: "clean",
	BuildDate: "2020-05-01T21:47:04Z",
	GoVersion: "go1.13.8b4",
	Compiler: "gc",
	Platform: "linux/amd64"
}

• Tekton Pipeline version:

Client version: 0.9.0
Pipeline version: v0.12.0

[bobcatfish]

Hey @chechiachang ! This seems like a reasonable request but I don't really have enough knowledge about how dashboard works to address it - would you be able to re-open this issue in https://github.com/tektoncd/dashboard instead? I think the folks over there could probably give you a better answer :D

[ghost]

I'm going to close this since it's an issue with the way the dashboard reruns pipelines. I don't think there's much that Pipelines can do to help here, though happy to discuss further if there is. Feel free to open an issue at https://github.com/tektoncd/dashboard if needed or reopen this one if you think there's more pipelines can do. Thanks for the issue report!