Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

YARN-11708: Setting maximum-application-lifetime using AQCv2 template… #7041

Merged
merged 9 commits into from
Oct 3, 2024
Merged

YARN-11708: Setting maximum-application-lifetime using AQCv2 template… #7041

Show file tree
Hide file tree
Changes from 1 commit
Commits
Show all changes
9 commits
Select commit Hold shift + click to select a range
12a584a
YARN-11708: Setting maximum-application-lifetime using AQCv2 template…
susheelgupta7 Sep 13, 2024
79788f1
adding getQueue to write lock
susheelgupta7 Sep 13, 2024
afdc38d
fixed all corner cases of queue ambiguity
susheelgupta7 Sep 17, 2024
51fa27d
added extra rmAppImpl arg
susheelgupta7 Sep 18, 2024
ea18cb6
changed the createqueue call to getorcreateQueueFromPlamemntContext
susheelgupta7 Sep 19, 2024
b0274bd
checkstyle fix
susheelgupta7 Sep 19, 2024
8c46ee7
used capacity scheduler for creating queue
susheelgupta7 Sep 20, 2024
3872a9b
Merging trunk to YARN-11708 to resolve conflict
susheelgupta7 Oct 2, 2024
d3f4ade
Again merging trunk to resolve conflict
susheelgupta7 Oct 2, 2024
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
21 changes: 19 additions & 2 deletions ...a/org/apache/hadoop/yarn/server/resourcemanager/scheduler/capacity/CapacityScheduler.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -3364,12 +3364,29 @@ public boolean moveReservedContainer(RMContainer toBeMovedContainer,
}
}

@Override
@Override
public long checkAndGetApplicationLifetime(String queueName,
susheelgupta7 marked this conversation as resolved.
Show resolved Hide resolved
long lifetimeRequestedByApp) {
CSQueue queue = getQueue(queueName);

// This handles the case where the queue does not exist,
// addressing the issue related to YARN-11708.
if (queue == null) {
QueuePath queuePath = new QueuePath(queueName);

writeLock.lock();
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This write lock should be before the get queue call, to properly handle the race condition, right?

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thanks for reviewing, yes it should be.

try {
queue = queueManager.createQueue(queuePath);
} catch (YarnException | IOException e) {
LOG.error("Failed to create queue '{}': ", queueName, e);
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Suggested change
LOG.error("Failed to create queue '{}': ", queueName, e);
LOG.error("Failed to create queue " + queueName, e);

Because the current log line wont log the exception, cause will think that is the 2nd param in the message

throw new YarnRuntimeException("Queue creation failed", e);
} finally {
writeLock.unlock();
}
}

readLock.lock();
try {
CSQueue queue = getQueue(queueName);
if (!(queue instanceof AbstractLeafQueue)) {
return lifetimeRequestedByApp;
}
Expand Down
55 changes: 55 additions & 0 deletions ...arn/server/resourcemanager/scheduler/capacity/TestCapacitySchedulerAutoQueueCreation.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -618,6 +618,61 @@ public void testAutoQueueCreationFailsForEmptyPathWithAQCAndWeightMode()
}
}

@Test
public void testAutoQueueCreationWithWeightModeAndMaxAppLifetimeFirstSubmittedApp()
throws Exception {
if (mockRM != null) {
mockRM.stop();
}

long maxRootLifetime = 20L;
long defaultRootLifetime = 10L;

QueuePath TEST_QUEUE = new QueuePath("root.test");

CapacitySchedulerConfiguration conf = setupSchedulerConfiguration();
conf.setQueues(ROOT, new String[] {"test"});
conf.setAutoQueueCreationV2Enabled(TEST_QUEUE, true);
conf.setCapacity(DEFAULT, "1w");
conf.setCapacity(TEST_QUEUE, "2w");
conf.setClass(YarnConfiguration.RM_SCHEDULER, CapacityScheduler.class,
ResourceScheduler.class);

conf.setMaximumLifetimePerQueue(ROOT, maxRootLifetime);
conf.setDefaultLifetimePerQueue(ROOT, defaultRootLifetime);

MockRM newMockRM = new MockRM(conf);
newMockRM.start();
((CapacityScheduler) newMockRM.getResourceScheduler()).start();

CapacityScheduler newCS =
(CapacityScheduler) newMockRM.getResourceScheduler();
newCS.checkAndGetApplicationLifetime("root.test.user", -1);

Assert.assertEquals(newCS.getMaximumApplicationLifetime("root.test.user"), 20L);

try {
Priority appPriority = Priority.newInstance(0);
RMApp app1 = MockRMAppSubmitter.submit(newMockRM,
MockRMAppSubmissionData.Builder.createWithMemory(1024, newMockRM)
.withAppPriority(appPriority)
.withQueue("root.test.user")
.build());

newMockRM.waitForState(app1.getApplicationId(), RMAppState.KILLED);
long totalTimeRun = app1.getFinishTime() - app1.getSubmitTime();

Assert.assertEquals(RMAppState.KILLED, app1.getState());
Assert.assertTrue("Application killed before default lifetime value",
totalTimeRun > (defaultRootLifetime * 1000));
Assert.assertTrue(
"Application killed after max lifetime value " + totalTimeRun,
totalTimeRun < (maxRootLifetime * 1000));
} finally {
((CapacityScheduler) newMockRM.getResourceScheduler()).stop();
newMockRM.stop();
}
}

/**
* This test case checks if a mapping rule can put an application to an auto
Expand Down
Loading