[jira] [Created] (IMPALA-7517) Hung scanner when soft memory limit exceeded

["Todd Lipcon (JIRA)" <ji...@apache.org>]

Todd Lipcon created IMPALA-7517:
-----------------------------------

             Summary: Hung scanner when soft memory limit exceeded
                 Key: IMPALA-7517
                 URL: https://issues.apache.org/jira/browse/IMPALA-7517
             Project: IMPALA
          Issue Type: Bug
          Components: Backend
    Affects Versions: Impala 3.1.0
            Reporter: Todd Lipcon
            Assignee: Todd Lipcon


As reported on the mailing list, this is a regression due to IMPALA-7096 (7ccf7369085aa49a8fc0daf6f91d97b8a3135682). The scanner thread has the following code:
 
    // Stop extra threads if we're over a soft limit in order to free up memory.
    if (!first_thread && mem_tracker_->AnyLimitExceeded(MemLimit::SOFT)) {
      break;
    }
 
    // Done with range and it completed successfully
    if (progress_.done()) {
      // All ranges are finished.  Indicate we are done.
      SetDone();
      break;
    }
 
    if (scan_range == nullptr && num_unqueued_files == 0) {
      unique_lock<mutex> l(lock_);
      // All ranges have been queued and DiskIoMgr has no more new ranges for this scan
      // node to process. This means that every range is either done or being processed by
      // another thread.
      all_ranges_started_ = true;
      break;
    }
  }
 
What if we have the following scenario:
 
T1) grab scan range 1 and start processing
 
T2) grab scan range 2 and start processing
 
T1) finish scan range 1 and see that 'progress_' is not done()
T1) loop around, get no scan range (there are no more), so set all_ranges_satrted_ and break
T1) thread exits
 
T2) finish scan range 2
T2) happen to hit a soft memory limit error due to pressure from other exec nodes, etc. Since we aren't the first thread, we break. (even though the first thread is no longer running)
T2) thread exits
 
Note that no one got to the point of calling SetDone() because we break due to the memory limit error _before_ checking progress_.Done().
 
Thus, the query will hang forever.



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)

---------------------------------------------------------------------
To unsubscribe, e-mail: issues-all-unsubscribe@impala.apache.org
For additional commands, e-mail: issues-all-help@impala.apache.org