[
https://issues.apache.org/jira/browse/HIVE-24723?focusedWorklogId=547221&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-547221
]
ASF GitHub Bot logged work on HIVE-24723:
-----------------------------------------
Author: ASF GitHub Bot
Created on: 03/Feb/21 18:55
Start Date: 03/Feb/21 18:55
Worklog Time Spent: 10m
Work Description: belugabehr commented on a change in pull request #1939:
URL: https://github.com/apache/hive/pull/1939#discussion_r569663760
##########
File path: ql/src/java/org/apache/hadoop/hive/ql/exec/tez/TezSessionPool.java
##########
@@ -395,20 +379,38 @@ int getInitialSize() {
} while (!deltaRemaining.compareAndSet(oldVal, oldVal + delta));
int toStart = oldVal + delta;
if (toStart <= 0) return createDummyFuture();
- LOG.info("Resizing the pool; adding " + toStart + " sessions");
-
- // 2) If we need to create some extra sessions, we'd do it just like
startup does.
- int threadCount = Math.max(1, Math.min(toStart,
- HiveConf.getIntVar(initConf,
ConfVars.HIVE_SERVER2_TEZ_SESSION_MAX_INIT_THREADS)));
- List<ListenableFutureTask<Boolean>> threadTasks = new
ArrayList<>(threadCount);
- // This is an async method, so always launch threads, even for a single
task.
- for (int i = 0; i < threadCount; ++i) {
- ListenableFutureTask<Boolean> task = ListenableFutureTask.create(
- new CreateSessionsRunnable(deltaRemaining));
- new Thread(task, "Tez pool resize " + i).start();
- threadTasks.add(task);
+ LOG.info("Resizing the pool; adding {} sessions", toStart);
+
+ final int threadCount =
+ Math.min(toStart, HiveConf.getIntVar(initConf,
ConfVars.HIVE_SERVER2_TEZ_SESSION_MAX_INIT_THREADS));
+
+ return createSessions(toStart, threadCount);
+ }
+
+ private ListenableFuture<List<Boolean>> createSessions(int sessionCount, int
maxParallel) {
+ Preconditions.checkArgument(sessionCount > 0);
+ Preconditions.checkArgument(maxParallel > 0);
Review comment:
This was previous checked in the calling methods. I have moved it here
to remove redundancy. Not an expensive operation and it's not called all that
often, so no worries even if it is doubled somewhere. Better safe than sorry
as the code changes.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
Issue Time Tracking
-------------------
Worklog Id: (was: 547221)
Time Spent: 1h 20m (was: 1h 10m)
> Use ExecutorService in TezSessionPool
> -------------------------------------
>
> Key: HIVE-24723
> URL: https://issues.apache.org/jira/browse/HIVE-24723
> Project: Hive
> Issue Type: Improvement
> Components: Tez
> Reporter: David Mollitor
> Assignee: David Mollitor
> Priority: Minor
> Labels: pull-request-available
> Time Spent: 1h 20m
> Remaining Estimate: 0h
>
> Currently there are some wonky home-made thread pooling action going on in
> {{TezSessionPool}}. Replace it with some JDK/Guava goodness.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
