it.getQueue() in /lockable-resources/. Reason: java.lang.NullPointerException about lockable-resources-plugin HOT 7 CLOSED

can you pls activate this logger and post here the log?

https://github.com/jenkinsci/lockable-resources-plugin/blob/08d4ee9edba86245b35d2a99e57a3b6c9ebcf238/src/main/java/org/jenkins/plugins/lockableresources/queue/QueuedContextStruct.java#L80C2-L80C2

thx

from lockable-resources-plugin.

We have the same problem. This is all I see (logging ALL of org.jenkins.plugins.lockableresources) after starting jenkins and loading the lockable resource page on Jenkins:

Jan 10, 2024 4:26:04 PM FINE org.jenkins.plugins.lockableresources.BackwardCompatibility compatibilityMigration
lockable-resources-plugin compatibility migration task run for 95 resources
Jan 10, 2024 4:26:04 PM INFO org.jenkins.plugins.lockableresources.NodesMirror createNodeResources
lockable-resources-plugin: configure node resources
Jan 10, 2024 4:26:04 PM FINE org.jenkins.plugins.lockableresources.FreeDeadJobs freePostMortemResources
lockable-resources-plugin free post mortem task run

I can reproduce it like this

• reserve all resources of a specific label
• start a pipeline job that tries to lock a resource with that label. This build ends up as a queuedContext in the XML
• restart Jenkins (https://jenkins/restart)

When Jenkins starts up again, and after opening the lockable resources page, I get the NullPointerException. Also when starting the same job again, it fails with that NullPointerException.

Pipeline: Groovy (workflow-cps) Version 3837.v305192405b_c0
Lockable Resources Version 1222.v3d55a_36d63e4
Jenkins 2.426.2

from lockable-resources-plugin.

feel free to review changes in the related PR #609

from lockable-resources-plugin.

I've tried the version of PR #609. Now the pipeline job waits for an available resource, but when I manually unreserve one of those resources, I still get the NullPointerException.

jenkins    | Caused by: java.lang.NullPointerException: Cannot invoke "org.jenkinsci.plugins.workflow.flow.FlowExecutionOwner.get()" because "this.executionRef" is null
jenkins    | 	at org.jenkinsci.plugins.workflow.cps.CpsStepContext.getExecution(CpsStepContext.java:217)
jenkins    | 	at org.jenkinsci.plugins.workflow.cps.CpsStepContext.getThreadGroupSynchronously(CpsStepContext.java:242)
jenkins    | 	at org.jenkinsci.plugins.workflow.cps.CpsStepContext.getThreadSynchronously(CpsStepContext.java:236)
jenkins    | 	at org.jenkinsci.plugins.workflow.cps.CpsStepContext.doGet(CpsStepContext.java:293)
jenkins    | 	at org.jenkinsci.plugins.workflow.support.DefaultStepContext.get(DefaultStepContext.java:75)
jenkins    | 	at org.jenkins.plugins.lockableresources.LockableResourcesManager.unreserve(LockableResourcesManager.java:961)
jenkins    | 	at org.jenkins.plugins.lockableresources.actions.LockableResourcesRootAction.doUnreserve(LockableResourcesRootAction.java:581)
jenkins    | 	at java.base/java.lang.invoke.MethodHandle.invokeWithArguments(MethodHandle.java:732)
jenkins    | 	at org.kohsuke.stapler.Function$MethodFunction.invoke(Function.java:397)
jenkins    | 	... 116 more

When I make the same change at https://github.com/mPokornyETM/lockable-resources-plugin/blob/fix-possible-npe-in-get-queue/src/main/java/org/jenkins/plugins/lockableresources/LockableResourcesManager.java#L965 (catch all exceptions), then I can unreserve the resource again. So then there is one resource free, but the pipeline only continues after unreserving another resource, as unreserving the first time only tries to continue with the invalid queuedContext.

from lockable-resources-plugin.

because "this.executionRef" is null

strange, the exception "because "this.executionRef" is null" happens in other plugin. Not here, I tries just to cacht it. May we chat samewhere else. I need maybe some more configs ... , Because I have really no idea, how can I simulate it.

Or can you try to write test case in this branch, which can simulate the scenario?

from lockable-resources-plugin.

Unfortunately, I can't reproduce it in a development environment, so I don't know what triggers the problem.

I can force it in a development environment by removing the "executionRef" element from a QueuedContextStruct in the XML. Not sure if this is useful for finding the actual problem though, but it seems to yield the same results in the lockable resources plugin.

    <org.jenkins.plugins.lockableresources.queue.QueuedContextStruct>
      <context class="org.jenkinsci.plugins.workflow.cps.CpsStepContext" plugin="[email protected]">
        <executionRef class="flow-owner" plugin="[email protected]">
          <job>lock-resource-long-time</job>
          <id>60</id>
        </executionRef>
        <id>7</id>
        <bodyHeads/>
        <body class="org.jenkinsci.plugins.workflow.cps.HandleBodyReference">
          <id>5</id>
        </body>
        <threadId>4</threadId>
        <stepDescriptorId>org.jenkins.plugins.lockableresources.LockStep</stepDescriptorId>
      </context>
      <lockableResourcesStruct>
      ...
    </org.jenkins.plugins.lockableresources.queue.QueuedContextStruct>

from lockable-resources-plugin.

Thats too much hacking. I mean, when the xml file is broken, or other plugin throws unexpected exceptions, it is too hard to handle it here.
Anyway, I thinkg the proposed changes are OK. Because it does not matter, why the item can not be proceeded. It shall be removed from queue.

from lockable-resources-plugin.