[YARN-10536] Client in distributedShell swallows interrupt exceptions

• Type: Bug
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Fix Version/s: 3.4.0, 3.3.1
• Component/s: client, distributed-shell
• Labels: pull-request-available
• Hadoop Flags: Reviewed
• Assignee: Ahmed Hussein Assign to me
• Reporter: Ahmed Hussein
• Created: 16/Dec/20 17:41
• Updated: 09/Jan/21 12:21
• Resolved: 17/Dec/20 23:14

Description

In applications.distributedshell.Client , the method monitorApplication loops waiting for the following conditions:

• Application fails: reaches YarnApplicationState.KILLED, or YarnApplicationState.FAILED
• Application succeeds: FinalApplicationStatus.SUCCEEDED or YarnApplicationState.FINISHED
• the time spent waiting is longer than clientTimeout (if it exists in the parameters).

When the Client thread is interrupted, it ignores the exception:

      // Check app status every 1 second.
      try {
        Thread.sleep(1000);
      } catch (InterruptedException e) {
        LOG.debug("Thread sleep in monitoring loop interrupted");
      }

Issue Links

• is depended upon by: YARN-10334 TestDistributedShell leaks resources on timeout/failure
• links to: GitHub Pull Request #2554