Troubleshoot Application Accelerator

This topic provides troubleshooting steps for development, accelerator authorship, and operations issues in Application Accelerator.

Installation issues

Depending on the error output, you can take the following actions to troubleshoot installation problems.

Verify installed packages

The package might be already installed. Verify this by running:

tanzu package installed list -n tap-install

Look for any package called accelerator.apps.tanzu.vmware.com.

Look at resource events

The error might be within the custom resources such as accelerator, Git repository, fragment, and so on. Find these errors by using the Kubernetes command line interface tool (kubectl).

Here is an example using the custom resource accelerator:

  1. Check for errors. For example, review the custom resource accelerator by running:

    kubectl get acc -n accelerator-system

    Note items in the output with a READY status False:

    NAME                       READY   REASON     AGE
appsso-starter-java        True    Ready      5h2m
where-for-dinner           True    Ready      5h2m
java-function              True    Ready      5h2m
java-rest-service          True    Ready      5h2m
java-server-side-ui        True    Ready      5h2m
node-express               True    Ready      5h2m
node-function              False   Not-Ready  5h2m
python-function            True    Ready      5h2m
spring-cloud-serverless    True    Ready      5h2m
spring-smtp-gateway        True    Ready      5h2m
tanzu-java-web-app         True    Ready      5h2m
tap-initialize             True    Ready      5h2m
weatherforecast-csharp     True    Ready      5h2m
weatherforecast-steeltoe   True    Ready      5h2m

  2. To see more information about any error events you found, run:

    kubectl get acc node-function -n accelerator-system -o yaml

    Look at the event section for more information about the error.

Development issues

Failure to generate a new project with URI is not absolute error

The generate command fails with the following error:

% tanzu accelerator generate test --server-url https://accelerator.example.com
Error: there was an error generating the accelerator, the server response was: "URI is not absolute"

Use:
  tanzu accelerator generate [flags]

Examples:
  tanzu accelerator generate <accelerator-name> --options '{"projectName":"test"}'

Flags:
  -h, --help                  help for generate
      --options string        options JSON string
      --options-file string   path to file containing options JSON string
      --output-dir string     directory that the zip file will be written to
      --server-url string     the URL for the Application Accelerator server

Global Flags:
      --context name      name of the kubeconfig context to use (default is current-context defined by kubeconfig)
      --kubeconfig file   kubeconfig file (default is $HOME/.kube/config)

there was an error generating the accelerator, the server response was: "URI is not absolute"

Error: exit status 1

✖  exit status 1

This indicates that the accelerator resource requested is not in a READY state. Review the instructions in the When Accelerator ready column is false section or contact your system admin.

Java project does not build in IntelliJ

Symptom:

When you create a Java project using the accelerator new project wizard in IntelliJ, it might not build correctly when first opened.

This issue mostly occurs in Maven projects. When you open the new project for the first time, a dialog box might appear in the bottom right side of IntelliJ asking you to Load Maven Project.

The Maven build scripts found dialog box with the Load Maven Project button.

Solution:

Click Load Maven Project and wait until the project builds.

If this does not fix the issue, delete the .idea folder and *.iml file and the reopen the project.

Accelerator authorship issues

General tips

Speed up the reconciliation of the accelerator

Set the git.interval to make the accelerator reconcile sooner. The default interval is 10 minutes, which is too long when developing an accelerator.

You can set this when using the YAML manifest:

apiVersion: accelerator.apps.tanzu.vmware.com/v1alpha1
kind: Accelerator
metadata:
  name: test-accelerator
spec:
  git:
    url: https://github.com/trisberg/test-accelerator
    ref:
      branch: main
    interval: 10s

You can also set this when creating the accelerator resource. To do so from the Tanzu CLI, run:

tanzu accelerator create test-accelerator --git-repo https://github.com/trisberg/test-accelerator --git-branch main --interval 10s

Use a source image with local accelerator source directory

You don’t have to use a Git repository when developing an accelerator. You can create an accelerator based on content in a local directory using --local-path when creating the accelerator resource.

Push the local path content to an OCI image by running:

tanzu accelerator create test-accelerator --local-path . --source-image REPO-PREFIX/test-accelerator --interval 10s

Where REPO-PREFIX is your own repository prefix. Use a repository that the deployed Application Accelerator system can access.

The interval is 10s so that you can push changes to the source-image repository and get faster reconcile time for the accelerator resource. When you have made changes to your accelerator source, push those changes by running:

tanzu accelerator push --local-path . --source-image REPO-PREFIX/test-accelerator

Where REPO-PREFIX is your own repository prefix. Use a repository that is accessible to the deployed Application Accelerator system.

Expression evaluation errors

Expression evaluation errors include:

  • Expression evaluated to null, such as:

    Could not read response from accelerator: java.lang.IllegalArgumentException: Expression '#mytestexp' evaluated to null

    In most cases, a typo in the variable name causes this error. Compare the expression with the defined options or any variables declared with let.

  • could not parse SpEL expression, such as:

    Could not read response from accelerator: Error reading manifest:could not parse SpEL expression at [Source: (InputStreamReader); line: 65, column: 1] (through reference chain: com.vmware.tanzu.accelerator.engine.manifest.Manifest["engine"]->com.vmware.tanzu.accelerator.engine.transform.transforms.Combo["let"]->java.util.ArrayList[0]->com.vmware.tanzu.accelerator.engine.transform.transforms.Let$DerivedSymbol["expression"])

    In most cases, an error in a let expression causes this error. Review the error message and, for more information, see SpEL samples.

  • SpelEvaluationException, such as:

    Could not read response from accelerator: org.springframework.expression.spel.SpelEvaluationException: EL1007E: Property or field 'test' cannot be found on null

    In most cases, an error in a transform expression causes this error. Review the error message and, for more information, see SpEL samples.

Operations issues

Accelerator persists in Tanzu Developer Portal after deletion

If an accelerator still displays in the Tanzu Developer Portal after it is deleted using the tanzu accelerator delete command, complete the following steps to delete:

  1. Navigate to your instance of the Tanzu Developer Portal.

  2. Search for the accelerator which should be deleted and select it.

  3. On the top right of the window, click the three dots, and select Unregister Template.

The accelerator is not longer displayed in the Tanzu Developer Portal Accelerator Catalog.

Check status of accelerator resources

Verify the status of accelerator resources by using kubectl or the Tanzu CLI:

  • From kubectl, run:

    kubectl get accelerators.accelerator.apps.tanzu.vmware.com -n accelerator-system

  • From the Tanzu CLI, run:

    tanzu accelerator list

Verify that the READY status is true for all accelerators.

When Accelerator ready column is blank

  1. View the status of accelerator-system by running:

    kubectl get deployment -n accelerator-system

    Example output:

    NAME                             READY   UP-TO-DATE   AVAILABLE   AGE
acc-engine                       1/1     1            1           3d5h
acc-server                       1/1     1            1           2d1h
accelerator-controller-manager   0/1     1            0           3d5h

  2. View the logs for any component with no Pods available by running:

    kubectl logs deployment/COMPONENT-NAME/ -n accelerator-system -p

    Where COMPONENT-NAME is the component with no pods you retrieved in the previous step.

    • If the log has the following error then the Flux CD source-controller is not installed:

      2021-11-18T20:55:18.963Z ERROR setup problem running manager {"error": "failed to wait for accelerator caches to sync: no matches for kind \"GitRepository\" in version \"source.toolkit.fluxcd.io/v1beta1\""}

    • If the log has the following error, the Tanzu Application Platform source-controller is not installed:

      2021-11-18T20:50:10.557Z ERROR setup problem running manager {"error": "failed to wait for accelerator caches to sync: no matches for kind \"ImageRepository\" in version \"source.apps.tanzu.vmware.com/v1alpha1\""}

When Accelerator ready column is false

View the REASON column for non-ready accelerators. Run:

kubectl get accelerators.accelerator.apps.tanzu.vmware.com -n accelerator-system

REASON: GitRepositoryResolutionFailed

For example:

$ kubectl get accelerators.accelerator.apps.tanzu.vmware.com -n accelerator-system
NAME        READY   REASON                             AGE
more-fun    False   GitRepositoryResolutionFailed      28s

  1. View the resource status. Run:

    kubectl get -oyaml accelerators.accelerator.apps.tanzu.vmware.com -n accelerator-system hello-fun

  2. Read status.conditions.message near the end of the output to learn the likely cause of failure. For example:

    status:
  address:
    url: http://accelerator-engine.accelerator-system.svc.cluster.local/invocations
  artifact:
    message: 'unable to clone ''https://github.com/vmware-tanzu/application-accelerator-samples'',
      error: couldn''t find remote ref "refs/heads/test"'
    ready: false
    url: ""
  conditions:
  - lastTransitionTime: "2021-11-18T21:05:47Z"
    message: |-
      failed to resolve GitRepository
      unable to clone 'https://github.com/vmware-tanzu/application-accelerator-samples', error: couldn't find remote ref "refs/heads/test"
    reason: GitRepositoryResolutionFailed
    status: "False"
    type: Ready
  description: Test-git
  observedGeneration: 1

    In this example, couldn't find remote ref "refs/heads/test" reveals that the branch or tag specified doesn’t exist.

    Another common problem is that the Git repository doesn’t exist. For example:

    status:
  address:
    url: http://accelerator-engine.accelerator-system.svc.cluster.local/invocations
  artifact:
    message: 'unable to clone ''https://github.com/vmware-tanzu/application-accelerator-sampl'',
      error: authentication required'
    ready: false
    url: ""
  conditions:
  - lastTransitionTime: "2021-11-18T21:09:52Z"
    message: |-
      failed to resolve GitRepository
      unable to clone 'https://github.com/vmware-tanzu/application-accelerator-sampl', error: authentication required
    reason: GitRepositoryResolutionFailed
    status: "False"
    type: Ready
  description: Test-git
  observedGeneration: 1

    An error message about failed authentication might display because the Git repository doesn’t exist. For example:

    unable to clone 'https://github.com/vmware-tanzu/application-accelerator-sampl', error: authentication required

REASON: GitRepositoryResolutionPending

For example:

$ kubectl get accelerators.accelerator.apps.tanzu.vmware.com -n accelerator-system
NAME        READY   REASON                             AGE
more-fun    False   GitRepositoryResolutionPending     28s

  1. See the resource status. Run:

    kubectl get -oyaml accelerators.accelerator.apps.tanzu.vmware.com -n accelerator-system hello-fun

  2. Locate status.conditions at the end of the output. For example:

    status:
  address:
    url: http://accelerator-engine.accelerator-system.svc.cluster.local/invocations
  artifact:
    message: ""
    ready: false
    url: ""
  conditions:
  - lastTransitionTime: "2021-11-18T20:17:38Z"
    message: GitRepository not yet resolved
    reason: GitRepositoryResolutionPending
    status: "False"
    type: Ready
  description: Test-git
  observedGeneration: 1

  3. Verify that the Flux system is running and that the READY column has 1/1. Run:

    kubectl get -n flux-system deployment/source-controller

    Example output:

    NAME                READY   UP-TO-DATE   AVAILABLE   AGE
source-controller   1/1     0            0           5d4h

REASON: ImageRepositoryResolutionPending

For example:

$ kubectl get accelerators.accelerator.apps.tanzu.vmware.com -n accelerator-system
NAME        READY   REASON                             AGE
more-fun    False   ImageRepositoryResolutionPending   28s

  1. See the resource status. Run:

    kubectl get -oyaml accelerators.accelerator.apps.tanzu.vmware.com -n accelerator-system hello-fun

  2. Locate status.conditions at the end of the output. For example:

    $ kubectl get -oyaml accelerators.accelerator.apps.tanzu.vmware.com -n accelerator-system more-fun

apiVersion: accelerator.apps.tanzu.vmware.com/v1alpha1
kind: Accelerator
metadata:
  annotations:
    kubectl.kubernetes.io/last-applied-configuration: |
      {"apiVersion":"accelerator.apps.tanzu.vmware.com/v1alpha1","kind":"Accelerator","metadata":{"annotations":{},"name":"more-fun","namespace":"accelerator-system"},"spec":{"description":"Test-image","source":{"image":"trisberg/more-fun-source"}}}
  creationTimestamp: "2021-11-18T20:32:36Z"
  generation: 1
  name: more-fun
  namespace: accelerator-system
  resourceVersion: "605401"
  uid: 407b565d-14aa-44fe-ad8d-c9b3c3a7e5ce
spec:
  description: Test-image
  source:
    image: trisberg/more-fun-source
status:
  address:
    url: http://accelerator-engine.accelerator-system.svc.cluster.local/invocations
  artifact:
    message: ""
    ready: false
    url: ""
  conditions:
  - lastTransitionTime: "2021-11-18T20:32:36Z"
    message: ImageRepository not yet resolved
    reason: ImageRepositoryResolutionPending
    status: "False"
    type: Ready
  description: Test-image
  observedGeneration: 1

  3. Verify that Tanzu Application Platform source-controller system is running and the READY column has 1/1. Run:

    kubectl get -n source-system deployment/source-controller-manager

    Expected output:

    NAME                        READY   UP-TO-DATE   AVAILABLE   AGE
source-controller-manager   1/1     0            0           5d5h
check-circle-line exclamation-circle-line close-line
Scroll to top icon