Was running a CloudBees Core 2.204.1.1 managed master. Restarted the master; I think there was a build running which had a node block with an unsatisfied label. At any rate, restart was very slow, and in the course of inspecting thread dumps during startup I noticed something which looked inappropriate:
"jenkins.util.Timer [#3]" ... waiting for monitor entry ...
java.lang.Thread.State: BLOCKED (on object monitor)
at jenkins.model.lazy.AbstractLazyLoadRunMap.getByNumber(AbstractLazyLoadRunMap.java:371)
- waiting to lock <...> (a hudson.model.RunMap)
at jenkins.model.lazy.LazyBuildMixIn.getBuildByNumber(LazyBuildMixIn.java:228)
at org.jenkinsci.plugins.workflow.job.WorkflowJob.getBuildByNumber(WorkflowJob.java:232)
at org.jenkinsci.plugins.workflow.job.WorkflowJob.getBuildByNumber(WorkflowJob.java:103)
at hudson.model.Run.fromExternalizableId(Run.java:2446)
at org.jenkinsci.plugins.workflow.support.steps.ExecutorStepExecution$PlaceholderTask.runForDisplay(ExecutorStepExecution.java:503)
at org.jenkinsci.plugins.workflow.support.steps.ExecutorStepExecution$PlaceholderTask.getOwnerTask(ExecutorStepExecution.java:436)
at org.jenkinsci.plugins.workflow.support.steps.ExecutorStepExecution$PlaceholderTask.getAffinityKey(ExecutorStepExecution.java:600)
at hudson.model.LoadBalancer$1.assignGreedily(LoadBalancer.java:119)
at hudson.model.LoadBalancer$1.map(LoadBalancer.java:107)
at hudson.model.LoadBalancer$2.map(LoadBalancer.java:167)
at hudson.model.Queue.maintain(Queue.java:1640)
at hudson.model.Queue$MaintainTask.doRun(Queue.java:2898)
at ...
Here getAffinityKey is waiting to load a build record (Run object) merely to determine the name of the owning job, which is silly since this information is available simply from the runId.
Bug
Minor
JENKINS-36547 Queue.Task.getFullDisplayName is a poor choice of key for LoadBalancer.CONSISTENT_HASH
