srdo commented on a change in pull request #2957: STORM-3335 allow timing out
when scheduling a topology
URL: https://github.com/apache/storm/pull/2957#discussion_r255291976
##########
File path:
storm-server/src/main/java/org/apache/storm/scheduler/resource/ResourceAwareScheduler.java
##########
@@ -136,10 +146,24 @@ private void scheduleTopology(TopologyDetails td,
Cluster cluster, final User to
return;
}
+ final IStrategy finalRasStrategy = rasStrategy;
for (int i = 0; i < maxSchedulingAttempts; i++) {
SingleTopologyCluster toSchedule = new
SingleTopologyCluster(workingState, td.getId());
try {
- SchedulingResult result = rasStrategy.schedule(toSchedule, td);
+ SchedulingResult result = null;
+ Future<SchedulingResult> schedulingFuture =
backgroundScheduling.submit(
+ () -> finalRasStrategy.schedule(toSchedule, td)
+ );
+ try {
+ result = schedulingFuture.get(schedulingTimeoutSeconds,
TimeUnit.SECONDS);
+ } catch (TimeoutException te) {
+ markFailedTopology(topologySubmitter, cluster, td,
"Scheduling took too long for "
+ + td.getId() + " using strategy " +
rasStrategy.getClass().getName() + " timeout after "
+ + schedulingTimeoutSeconds + " seconds.");
+ schedulingFuture.cancel(true);
+ rasStrategy.stop();
Review comment:
Consider dropping the `stop` method and make the scheduler check for
interrupts instead. You're already sending an interrupt to it via the `cancel`
call above.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
[email protected]
With regards,
Apache Git Services
