Uploaded image for project: 'WildFly Core'
  1. WildFly Core
  2. WFCORE-6157

Investigate what's wrong with operations cancellation


    • Icon: Bug Bug
    • Resolution: Done
    • Icon: Major Major
    • 20.0.0.Final
    • None
    • Server
    • None

      WFCORE-6156 fix of ContainerStateMonitor led to these failures in WildFly Core test suite.

      [ERROR] Failures:
      [ERROR] SecondaryReconnectTestCase.test01_OrderedExtensionsAndDeployments:112->testReconnect:184 expected:<[ok]> but was:<[restart-required]>
      [ERROR] CoreResourceManagementTestCase.testCanFindServerRestartRequiredAfterChangingSocketBindingPortOffset:860 expected:<re[load]-required> but was:<re[start]-required>
      [ERROR] ServerRestartRequiredTestCase.testChangeGroup:211->assertServerState:371 expected:<r[unning]> but was:<r[estart-required]>
      [ERROR] Errors:
      [ERROR] SecondaryReconnectTestCase.test03_DeploymentOverlays:131->testReconnect:145->createServer:273 » MgmtOperation
      [ERROR] IgnoredResourcesProfileCloneTestCase.afterClass:91 » Timeout Managed servers w...
      [ERROR] Tests run: 582, Failures: 3, Errors: 2, Skipped: 20

      During investigation of this regression I noticed the awaitStability() was interrupted (in OperationContextImpl.releaseStepLocks() method) and left server in inconsistent state.
      That is the root cause of these regressions.

      I decided to forbid thread interruption of a thread awaiting container stability.
      We will need to revisit if this feature is a MUST HAVE or we can live without it.
      Notice it is not possible to block forever when awaitStability() is used with timeout configured.

            bstansbe@redhat.com Brian Stansberry
            ropalka Richard Opalka
            0 Vote for this issue
            1 Start watching this issue