I have a below pipeline

apiVersion: tekton.dev/v1beta1
kind: Pipeline
metadata:
  name: git-clone-pipeline
spec:
  params:
    - name: repo-url
      type: string
  workspaces:
    - name: shared-workspace
  tasks:
    - name: clone-repository
      taskRef:
        name: git-clone
      workspaces:
        - name: output
          workspace: shared-workspace
      params:
        - name: url
          value: "$(params.repo-url)"
        - name: deleteExisting
          value: "true"
    - name: build
      taskRef:
        name: gradle
      runAfter:
        - "clone-repository"
      params:
        - name: TASKS
          value: build
        - name: GRADLE_IMAGE
          value: docker.io/library/gradle:jdk17-alpine@sha256:dd16ae381eed88d2b33f977b504fb37456e553a1b9c62100b8811e4d8dec99ff
        - name: PROJECT_DIR
          value: ./discount-api
      workspaces:
        - name: source
          workspace: shared-workspace

And pipeline-run

apiVersion: tekton.dev/v1beta1
kind: PipelineRun
metadata:
  name: run-pipeline
  namespace: tekton-pipelines
spec:
  serviceAccountName: git-service-account
  pipelineRef:
    name: git-clone-pipeline
  workspaces:
    - name: shared-workspace
      emptyDir: {}
  params:
    - name: repo-url
      value: [email protected]:anandjaisy/discount.git

And a project structure directory as

Facing an issue during the pipeline run

2022-06-18T08:17:57.775506334Z Directory '/workspace/source/discount-api' does not contain a Gradle build.

This issue is related to the file not being found. The git-clone task has cloned the code somewhere in the cluster. How do I know where is the code?

kubectl get pods

run-pipeline-build-pod                               0/1     Error       0          173m
run-pipeline-fetch-source-pod                        0/1     Completed   0          173m
tekton-dashboard-b7b8599c6-wf7b2                     1/1     Running     0          12d
tekton-pipelines-controller-674dd45d79-529pc         1/1     Running     0          12d
tekton-pipelines-webhook-86b8b9d87b-qmxzk            1/1     Running     0          12d
tekton-triggers-controller-6d769dddf7-847nt          1/1     Running     0          12d
tekton-triggers-core-interceptors-69c47c4bb7-77bvt   1/1     Running     0          12d
tekton-triggers-webhook-7c4fc7c74-lgm79              1/1     Running     0          12d

If we do logs on POD run-pipeline-build-pod gets an exception as above.

kubectl exec --stdin --tty <Pod-name> -- /bin/bash

The above command is used to go inside the pod, however, in the above case the run-pipeline-fetch-source-pod is completed and we can't apply the above code.

How can I solve this file issue ?

CodePudding user response：

You're using an emptyDir as a workspace. You should use a PVC. You could fix your pipelinerun with something like this:

workspaces:
- name: shared-workspace
  volumeClaimTemplate:
    spec:
      accessModes:
      - ReadWriteOnce
      resources:
        requests:
          storage: 2Gi

The workspace is used by your git clone task. Describing the git-clone pod from your pipelinerun, you should find a volume, of type emptyDir. Which is where data is cloned.

Then, using that workspace in another Task, being an emptyDir, in another Pod: you're starting from scratch, there's no data when pod starts. A PVC would allow you to share data in between tasks.

---

As for debugging: once pods are completed, there's not much to do.

Debugging pipelineruns, if it exits too fast for me to kubectl exec -it a Pod, I would try to change the failing container arguments (editing the corresponding Task), adding some "sleep 86400" or something equivalent, instead of its usual command. Then re-start a pipelinerun.

While in your case, if you already had a PVC as your workspace: it would have been easier to just start a Pod (create a deployment), attaching the PVC created by your pipelinerun, checking what's in there, ...