TisonKun commented on a change in pull request #10048: 
[FLINK-14462][coordination] Remove JobGraph#allowQueuedScheduling flag because 
it is always true
URL: https://github.com/apache/flink/pull/10048#discussion_r341891476
 
 

 ##########
 File path: 
flink-runtime/src/test/java/org/apache/flink/runtime/jobmanager/scheduler/SchedulerIsolatedTasksTest.java
 ##########
 @@ -212,14 +154,12 @@ public void testScheduleWithDyingInstances() throws 
Exception {
 
                // cannot get another slot, since all instances are dead
                try {
-                       testingSlotProvider.allocateSlot(new 
ScheduledUnit(getDummyTask()), false, SlotProfile.noRequirements(), 
TestingUtils.infiniteTime()).get();
+                       testingSlotProvider
+                               .allocateSlot(new 
ScheduledUnit(getDummyTask()), SlotProfile.noRequirements(), 
TestingUtils.infiniteTime())
+                               .get(2000L, TimeUnit.MILLISECONDS);
 
 Review comment:
   I verify with the diff below
   
   ```diff
   @@ -155,11 +155,11 @@ public class SchedulerIsolatedTasksTest extends 
SchedulerTestBase {
                   // cannot get another slot, since all instances are dead
                   try {
                           testingSlotProvider
   -                               .allocateSlot(new 
ScheduledUnit(getDummyTask()), SlotProfile.noRequirements(), 
TestingUtils.infiniteTime())
   -                               .get(2000L, TimeUnit.MILLISECONDS);
   +                               .allocateSlot(new 
ScheduledUnit(getDummyTask()), SlotProfile.noRequirements(), 
Time.milliseconds(0L))
   +                               .get();
                           fail("Scheduler served a slot from a dead instance");
   -               } catch (TimeoutException e) {
   -                       // expected
   +               } catch (Exception e) {
   +                       assertTrue(e.getCause() instanceof TimeoutException);
                   }
    
                   // now the latest, the scheduler should have noticed 
(through the lazy mechanisms)
   ```
   
   However, due to we `get` in main thread and schedule timeout future in main 
thread, we dead lock by main thread itself...Maybe we introduce a new test with 
`ComponentMainThreadExecutorServiceAdapter.forSingleThreadExecutor`. I will 
appreciate it if you can help port this test.

----------------------------------------------------------------
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


With regards,
Apache Git Services

Reply via email to