Cannot scale jira on aks

Question

Trying to host Jira on AKS using helm repo https://atlassian.github.io/data-center-helm-charts and chart atlassian-data-center/jira, azure file share is used as Persistent Volume.

Below is values.yaml used in installation.

replicaCount: 1
serviceAccount:
  create: true
  name: jira
database:
  type: mssql
  url: jdbc:******
  driver: com.microsoft.sqlserver.jdbc.SQLServerDriver
  credentials:
    secretName: jira-database-credentials
    usernameSecretKey: ******
    passwordSecretKey: ******
volumes:
  localHome:
    customVolume:
      persistentVolumeClaim:
        claimName: local-pvc
  sharedHome:
    customVolume:
      persistentVolumeClaim:
        claimName: shared-pvc   
    nfsPermissionFixer:
      enabled: true
      mountPath: "/shared-home"
ingress:
  create: true
  className: "nginx"
  nginx: true
  maxBodySize: 250m
  host: "host"
  path: "/"
  https: true
  tlsSecretName: "tlsSecretName"
additionalFiles:
  - name: jira-configs
    type: configMap
    key: cluster.properties
    mountPath: /var/atlassian/application-data/jira

jira:
  readinessProbe:
    initialDelaySeconds: 300
    periodSeconds: 10
    timeoutSeconds: 5
    failureThreshold: 6
    successThreshold: 1
  startupProbe:
    initialDelaySeconds: 300
    periodSeconds: 10
    timeoutSeconds: 5
    failureThreshold: 6
    successThreshold: 1
  livenessProbe:
    enabled: false
  resources:
    jvm:
      maxHeap: "8G"
    container:
      requests:
        cpu: "2"
        memory: "4G"
clustering:
  enabled: true

When there is single pod running, I can see cache.lock file successfully created but as soon as I try to scale, I see file is getting marked for deletion and throwing exception mentioned below

ERROR: Error creating bundle cache.
java.lang.Exception: Unable to create bundle cache lock file: java.nio.file.NoSuchFileException: /var/atlassian/application-data/jira/plugins/.osgi-plugins/felix/felix-cache/cache.lock
        at org.apache.felix.framework.cache.BundleCache.<init>(BundleCache.java:161)

It appears Jira is trying to delete the cache.lock file while it's already in use or locked by some other process. I tried to delete file manually by going inside the pod and can see same behavior.

In Persistent volumes I have set the mount options as below, tried with accessModes ReadWriteOnce and ReadWriteMany but same error.

mountOptions:
    - dir_mode=0777
    - file_mode=0777
    - uid=2001
    - gid=2001
    - mfsymlinks
    - cache=strict
    - nosharesock
    - nobrl

Any help is much appreciated.

score 0 · Answer 1 · answered Jun 30 '23 at 13:07

After looking at the logs, I found that each share is being locked for exclusive use by pod so I provisioned share dynamically and issue got resolved.

Update volume section in values.yaml file looks like below.

volumes:
  localHome:
    persistentVolumeClaim:
      create: true
      storageClassName: private-azurefile-csi
      accessModes:
        - ReadWriteOnce
      resources:
        requests:
          storage: 50Gi

Cannot scale jira on aks

1 Answers1