Plugin container accessor is null

[Chris Butwell]

Running the 1.1.16 TFS jar. Suddenly today this error occurs. Recycled - but didn't fix. Then tried the 1.1.17 hotfix you created for us but that did not fix (didn't think it would but decided to try).

Note we use remote agents. Main server is Linux, then two Windows Server hosts, Mac hosts, and linux hosts.

com.atlassian.bamboo.repository.RepositoryException: com.atlassian.util.concurrent.LazyReference$InitializationException: java.lang.NullPointerException: Plugin container accessor is null. Plugin: com.stellarity.bamboo.tfs-repository-plugin. Module name: com.stellarity.bamboo.repository.TfsRepository.
    at com.atlassian.bamboo.v2.trigger.DefaultChangeDetectionManager.collectChangesSinceLastBuildInternal(DefaultChangeDetectionManager.java:367)
    at com.atlassian.bamboo.v2.trigger.DefaultChangeDetectionManager.collectChangesSinceRevisions(DefaultChangeDetectionManager.java:307)
    at com.atlassian.bamboo.v2.trigger.DefaultChangeDetectionManager.collectChangesSinceRevisions(DefaultChangeDetectionManager.java:192)
    at com.atlassian.bamboo.v2.trigger.DefaultChangeDetectionManager.collectChangesSinceLastBuildIfTriggered(DefaultChangeDetectionManager.java:130)
    at com.atlassian.bamboo.v2.trigger.ChangeDetectionListenerAction.testIfBuildShouldStart(ChangeDetectionListenerAction.java:116)
    at com.atlassian.bamboo.plan.PlanExecutionManagerImpl$4.call(PlanExecutionManagerImpl.java:522)
    at com.atlassian.bamboo.plan.PlanExecutionManagerImpl$4.call(PlanExecutionManagerImpl.java:505)
    at com.atlassian.util.concurrent.ManagedLocks$ManagedLockImpl.withLock(ManagedLocks.java:312)
    at com.atlassian.bamboo.plan.PlanExecutionLockServiceImpl.lock(PlanExecutionLockServiceImpl.java:85)
    at com.atlassian.bamboo.plan.PlanExecutionManagerImpl.doWithProcessLock(PlanExecutionManagerImpl.java:796)
    at com.atlassian.bamboo.plan.PlanExecutionManagerImpl.startConditionalBuild(PlanExecutionManagerImpl.java:504)
    at com.atlassian.bamboo.plan.PlanExecutionManagerImpl.start(PlanExecutionManagerImpl.java:578)
    at com.atlassian.bamboo.plan.PlanExecutionManagerImpl.start(PlanExecutionManagerImpl.java:595)
    at com.atlassian.bamboo.plan.DelegatingPlanExecutionManager.start(DelegatingPlanExecutionManager.java:95)
    at com.atlassian.bamboo.plan.NonBlockingPlanExecutionServiceImpl$4.startPlanExecution(NonBlockingPlanExecutionServiceImpl.java:240)
    at com.atlassian.bamboo.plan.NonBlockingPlanExecutionServiceImpl$4.lambda$call$130(NonBlockingPlanExecutionServiceImpl.java:219)
    at com.atlassian.bamboo.plan.NonBlockingPlanExecutionServiceImpl$4$$Lambda$97/1530148903.run(Unknown Source)

[Sergey Podobry]

Hi Chris,

The plugin was not loaded by the main server. I need to see the Bamboo log from its startup. Could you attach it here?

[Chris Butwell]

Let me see if I can scrub the log file so security will let me post. Work at a financial organization and logs are very restricted.

What is interesting is we were running for about 2 weeks and then suddenly this occurred, so it was not the result of a recycle. As part of trying to solve this issue, we did recycle Bamboo. Now that we have recycled, all build plans with a TFS repository still fail, however we found a "fix". We can now add a new repository (before the recycle would get null pointer trying to add a TFS repo) to the build plan, then delete that newly added repository and then the previously configured TFS repo begins working again.

[Sergey Podobry]

Search the log for any ERROR or WARN messages.

[Sergey Podobry]

Your description looks like there are 2 plugin JAR files in the $BAMBOO_HOME/plugins directory. Could you check it?

[Chris Butwell]

Just one TFS plugin in the $BAMBOO_HOME/plugins directory. Error started with the 1.1.16 jar we removed that jar, uploaded 1.1.17 and restarted. Problems continue. Atlassian is also looking at this issue currently. Not clear on the cause but we are now ok with GIT based repositories but TFS projects continue to have problems. I'll keep you posted on their replies. Right now they are analyzing VCS_LOCATION data as it appears to have issues.

[Sergey Podobry]

Ok. Did you find anything in the log?

[Sergey Podobry]

Chris, do you have any news? Did Atlassian guys find anything?

[Sergey Podobry]

We believe this is a corruption of the plugin cache. In this issue (https://community.atlassian.com/t5/Bamboo-questions/Remote-Agent-gitv2-not-available/qaq-p/858832) a built-in git plugin failed to load and cache cleanup solved the issue.