Uploaded image for project: 'Jenkins'
  1. Jenkins
  2. JENKINS-72650

"Periodic Jenkins queue maintenance" thread can soft-lock Jenkins queue while waiting for LDAP blocking search


    • Icon: Bug Bug
    • Resolution: Unresolved
    • Icon: Minor Minor
    • ldap-plugin
    • None

      We recently noticed that our whole Jenkins instance would go completely unresponsive for dozens of minutes, meaning all jobs get frozen.


      During these episodes, the Jenkins log is filled with the following messages

      20240201 09:46:37.511+0000 [id={}}}{{{}172{}}}{{{}]   INFO    o.j.p.w.s.concurrent.Timeout#lambda$ping$0: Running CpsFlowExecution[Owner[<REDACTED>{}}}{{{}]] unresponsive for 12 min


      If we pause the threads and review the stacks, we can notice the following:

      • All unresponsive threads are blocked waiting for the job queue (see Running CpsFlowExecution.txt)
      • The thread Periodic Jenkins queue maintenance has the lock on the job queue (see AtmostOneTaskExecutor.txt)
      • This thread is blocked waiting for a reply from the LDAP server


      The culprit is probably the LDAP server, but it doesn't seems robust that LDAP queries don't have a configured timeout. According to Oracle's documentation, one could set the timeout using the property com.sun.jndi.ldap.connect.timeout.

            Unassigned Unassigned
            tdubouch Thomas Duboucher
            0 Vote for this issue
            1 Start watching this issue