Closed tam512 closed 4 years ago
My first reaction without reproducing this is that this is a tekton dashboard issue. I don't think we actually run the build-push step. It would have blown up with lots of errors if it did. So I don't think we have a major issue here. I have only seen 2 states for steps in the dashboard. The red cross or the green check. Not sure if there is a 3rd one representing didn't execute. We will investigate it a bit more today to see how it looks like from the command line view today and decide on the next steps.
We were able to re-produce this issue as below
github.ibm.com
from the tekton dashboard.compile time errors
, so the build-bud
step fails.Results:
As the pipelinerun failed, the tekton dashboard shows exact screenshot as per this issue.
However when we investigated from the oc logs command we found below results
oc logs -f my-java-mp-webhook-1571152644-build-task-wndcp-pod-2f0e7d -c step-build-bud
Logs Snippet for build-bud
container in the pod (which was as expected to build fail):-
Downloaded from central: https://repo.maven.apache.org/maven2/com/google/collections/google-collections/1.0/google-collections-1.0.jar (640 kB at 1.2 MB/s)
[INFO] Changes detected - recompiling the module!
[INFO] Compiling 3 source files to /project/user-app/target/classes
[INFO] -------------------------------------------------------------
[ERROR] COMPILATION ERROR :
[INFO] -------------------------------------------------------------
[ERROR] /project/user-app/src/main/java/dev/appsody/starter/StarterApplication.java:[8,5] <identifier> expected
[ERROR] /project/user-app/src/main/java/dev/appsody/starter/StarterApplication.java:[9,2] reached end of file while parsing
[INFO] 2 errors
[INFO] -------------------------------------------------------------
[INFO] ------------------------------------------------------------------------
[INFO] BUILD FAILURE
[INFO] ------------------------------------------------------------------------
[INFO] Total time: 57.022 s
[INFO] Finished at: 2019-10-15T15:27:33Z
[INFO] ------------------------------------------------------------------------
[ERROR] Failed to execute goal org.apache.maven.plugins:maven-compiler-plugin:3.1:compile (default-compile) on project starter-app: Compilation failure: Compilation failure:
[ERROR] /project/user-app/src/main/java/dev/appsody/starter/StarterApplication.java:[8,5] <identifier> expected
[ERROR] /project/user-app/src/main/java/dev/appsody/starter/StarterApplication.java:[9,2] reached end of file while parsing
[ERROR] -> [Help 1]
[ERROR]
[ERROR] To see the full stack trace of the errors, re-run Maven with the -e switch.
[ERROR] Re-run Maven using the -X switch to enable full debug logging.
[ERROR]
[ERROR] For more information about the errors and possible solutions, please read the following articles:
[ERROR] [Help 1] http://cwiki.apache.org/confluence/display/MAVEN/MojoFailureException
subprocess exited with status 1
subprocess exited with status 1
error building at STEP "RUN mvn install -DskipTests": exit status 1
oc logs -f my-java-mp-webhook-1571152644-build-task-wndcp-pod-2f0e7d -c step-build-push
Logs never returned anything :
[root@crofts1 common]# oc logs -f my-java-mp-webhook-1571152644-build-task-wndcp-pod-2f0e7d -c step-build-push
[root@crofts1 common]#
Observation:
This means that build-bud
container step failed, and the build-push
never started. However the tekton-dashboard shows the build-push
step got completed. So we think it is a Tekton Dashboard issue.
This issue is marked as blocked by tekton and we have opened the issue in Tekton Dashboard repository:- https://github.com/tektoncd/dashboard/issues/644
Unlikely that Tekton would address in time to make release 0.2.0 and I removed the tag. We should pursue the issue under the tekton dashboard repo.
@aadeshpa @stephenkinder it looks like Tekton fixed their issue. Will this be making it into the next Kabanero release?
@gecock from what In understand, this fix in only in tekton 0.8, which we are not picking up in kabanero 0.3.0. We can look into pulling this in for 0.4.0.
I don't think we are going to be able to contain pulling in tekton 0.8 in 0.4.0 with the time we have lost working on 0.3.0 issues in 0.4.0. Might have to wait another iteration for this.
@mtamboli - Are you good if I update the Kabanero release to v0.5.0 based on the comments provided by Vijai above?
@kvijai82 - Is Kabanero 0.5.0 pulling tekton 0.8?
As far as I know the plan is to pull in tekton 0.8 into Kabanero 0.5.0. We were able to successfully test triggers.
I verified with collections 0.4.0 release with below setup
collections:
repositories:
- activateDefaultCollections: true
name: central
url: https://github.com/kabanero-io/collections/releases/download/0.4.0/kabanero-index.yaml
github: {}
landing: {}
tekton: {}
version: 0.4.0
webhook: {}
status:
admissionControllerWebhook:
ready: "True"
appsody:
ready: "True"
version: 0.2.2
cli:
hostnames:
- kabanero-cli-kabanero.apps.vespers.os.fyre.ibm.com
ready: "True"
collectionController:
ready: "True"
version: 0.4.0
kabaneroInstance:
ready: "True"
version: 0.4.0
knativeEventing:
ready: "True"
version: 0.9.0
landing:
ready: "True"
version: 0.4.0
serverless:
knativeServing:
ready: "True"
version: 0.8.1
ready: "True"
version: 1.1.0
tekton:
ready: "True"
version: v0.7.0
My Tekton-dashboard version is v0.2.1
[root@vespers-inf auth]# kubectl get pod tekton-dashboard-6c4688f8bb-htmwc -n tekton-pipelines -o yaml | grep image
image: openshift/oauth-proxy@sha256:6bc1759a3202b4614739f12441461e344907f6b3f758c34314284debe36d4e15
imagePullPolicy: IfNotPresent
image: gcr.io/tekton-releases/github.com/tektoncd/dashboard/cmd/dashboard@sha256:fa7cef8ad3cbb1d79af45d87ec7e8913f886b3d55b10acd55995f907aced9ebb
imagePullPolicy: IfNotPresent
imagePullSecrets:
image: docker.io/openshift/oauth-proxy@sha256:6bc1759a3202b4614739f12441461e344907f6b3f758c34314284debe36d4e15
imageID: docker.io/openshift/oauth-proxy@sha256:6bc1759a3202b4614739f12441461e344907f6b3f758c34314284debe36d4e15
image: gcr.io/tekton-releases/github.com/tektoncd/dashboard/cmd/dashboard@sha256:fa7cef8ad3cbb1d79af45d87ec7e8913f886b3d55b10acd55995f907aced9ebb
imageID: gcr.io/tekton-releases/github.com/tektoncd/dashboard/cmd/dashboard@sha256:fa7cef8ad3cbb1d79af45d87ec7e8913f886b3d55b10acd55995f907aced9ebb
This shows if a step fails the pipelinerun in dashboard does not go to next step.
@tam512 - According to Ashish the issue is not reproducible, can you please test it with the latest Tekton versions?
I'm testing with Kabanero 0.5.0-rc2 which includes Tekton version v0.8.0. Inject an error to pom.xml to make build step fails and I still see the later steps being green/pass
# oc get kabanero -n kabanero -o yaml
apiVersion: v1
items:
- apiVersion: kabanero.io/v1alpha1
kind: Kabanero
metadata:
annotations:
kubectl.kubernetes.io/last-applied-configuration: |
{"apiVersion":"kabanero.io/v1alpha1","kind":"Kabanero","metadata":{"annotations":{},"name":"kabanero","namespace":"kabanero"},"spec":{"collections":{"repositories":[{"activateDefaultCollections":true,"name":"central","url":"https://github.com/kabanero-io/collections/releases/download/0.5.0-rc.3/kabanero-index.yaml"}]},"version":"0.5.0"}}
creationTimestamp: "2020-01-14T22:56:07Z"
finalizers:
- kabanero.io.kabanero-operator
generation: 3
name: kabanero
namespace: kabanero
resourceVersion: "93188"
selfLink: /apis/kabanero.io/v1alpha1/namespaces/kabanero/kabaneros/kabanero
uid: 0c8b422a-3721-11ea-b529-00000a1a02d4
spec:
admissionControllerWebhook: {}
che:
cheOperatorInstance: {}
enable: false
kabaneroChe: {}
cliServices: {}
collectionController: {}
collections:
repositories:
- activateDefaultCollections: true
name: central
url: https://github.com/kabanero-io/collections/releases/download/0.5.0-rc.3/kabanero-index.yaml
events: {}
github: {}
landing: {}
targetNamespaces:
- svt-project1
- svt-project2
- svt-project3
tekton: {}
version: 0.5.0
status:
admissionControllerWebhook:
ready: "True"
appsody:
ready: "True"
version: 0.3.0
cli:
hostnames:
- kabanero-cli-kabanero.apps.toast.os.fyre.ibm.com
ready: "True"
collectionController:
ready: "True"
version: 0.5.0-rc.1
kabaneroInstance:
ready: "True"
version: 0.5.0
knativeEventing:
ready: "True"
version: 0.9.0
landing:
ready: "True"
version: 0.5.0-rc.1
serverless:
knativeServing:
ready: "True"
version: 0.9.0
ready: "True"
version: 1.2.0
tekton:
ready: "True"
version: v0.8.0
kind: List
metadata:
resourceVersion: ""
selfLink: ""
# kubectl get pod tekton-dashboard-5646944d99-bdfw2 -n tekton-pipelines -o yaml | grep image
image: openshift/oauth-proxy@sha256:6bc1759a3202b4614739f12441461e344907f6b3f758c34314284debe36d4e15
imagePullPolicy: IfNotPresent
image: gcr.io/tekton-releases/github.com/tektoncd/dashboard/cmd/dashboard@sha256:7ccc937d9dbd3a4e769ed81ab8256e369488813554700889d74855f811f161b0
imagePullPolicy: IfNotPresent
imagePullSecrets:
image: docker.io/openshift/oauth-proxy@sha256:6bc1759a3202b4614739f12441461e344907f6b3f758c34314284debe36d4e15
imageID: docker.io/openshift/oauth-proxy@sha256:6bc1759a3202b4614739f12441461e344907f6b3f758c34314284debe36d4e15
image: gcr.io/tekton-releases/github.com/tektoncd/dashboard/cmd/dashboard@sha256:7ccc937d9dbd3a4e769ed81ab8256e369488813554700889d74855f811f161b0
imageID: gcr.io/tekton-releases/github.com/tektoncd/dashboard/cmd/dashboard@sha256:7ccc937d9dbd3a4e769ed81ab8256e369488813554700889d74855f811f161b0
Container classification
Digest
sha256:7ccc937d9dbd3a4e769ed81ab8256e369488813554700889d74855f811f161b0
Tags
v0.3.0
Repository
github.com/tektoncd/dashboard/cmd/dashboard
Project
tekton-releases
@tam512 I just verified that I also see the bug when running Tekton Dashboard v0.3.0 and Tekton Pipelines v0.8.0.
This bug was fixed in the Tekton Dashboard v0.2.1 with PR https://github.com/tektoncd/dashboard/pull/653. However, the fix PR was reverted for Dashboard v0.3.0 when Tekton Pipelines v0.9.0 fixed an underlying issue causing the bug https://github.com/tektoncd/dashboard/pull/795.
Unfortunately, Dashboard v0.3.0 is now bundled with Tekton Pipelines v0.8.0. So, we have the bug again.
I just verified that once the version of Tekton Pipelines is bumped up to v0.9.0 or greater, this bug will be resolved:
UPDATE: I was assuming that the next Kabanero release would bump up the Tekton Pipelines version to v0.9.0 or greater. However, apparently that's not the case. So it might be worthwhile to revert https://github.com/tektoncd/dashboard/pull/795 to patch up the Dashboard to work with Tekton Pipelines v0.8.0.
This was fixed in tekton 0.7 and the fix was successfully tested. The bug showed up again when we pulled tekton 0.8. Opening new tracking issue for it.
@kvijai82 please post the new issue number here once you have it. Thanks!
Using Kabanero 0.2.0 rc1, running a java-microprofile-build-deploy-pipeline in tekton dashboard. There was a problem in the pom.xml which caused the build-bud step fails (red mark), but it continues to go on to the next step build-push and show as green completed with
No log available
1) If a build step fails, it should not go on to the next build step. 2) If it has to go on to the next step, then it should be failed also