[
https://issues.apache.org/jira/browse/YARN-6210?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15878818#comment-15878818
]
Daniel Templeton commented on YARN-6210:
----------------------------------------
Looks good. A couple tiny nits:
* {{FSAppAttempt.isStarved(Resource usage, Resource share)}}: probably a bad
name. It doesn't have anything to do with starving. It's a comparator.
* {quote}The reservation-at-lower-priority assert is retained.{quote} Nope,
the assignment-at-lower-priority assert is retained. The original test also
had an assert to test the reservation itself: {code} // Reserved container
should still be at lower priority
for (RMContainer container : app.getReservedContainers()) {
assertEquals(2,
container.getReservedSchedulerKey().getPriority().getPriority());
}{code}
* On the assert messages, please keep in mind what the person triaging the
failures will see, e.g. "Basic allocation failed expected:<1> but was:<0>."
The message doesn't explain the numbers. Not a major issue, and not something
I expect you to fix, but I wanted to point it out.
> FS: Node reservations can interfere with preemption
> ---------------------------------------------------
>
> Key: YARN-6210
> URL: https://issues.apache.org/jira/browse/YARN-6210
> Project: Hadoop YARN
> Issue Type: Bug
> Components: fairscheduler
> Affects Versions: 2.9.0
> Reporter: Karthik Kambatla
> Assignee: Karthik Kambatla
> Attachments: YARN-6210.1.patch, YARN-6210.2.patch, YARN-6210.3.patch
>
>
> Today, on a saturated cluster, apps with pending demand reserve nodes. A new
> app might not be able to preempt resources because these nodes are already
> reserved. This can be reproduced by the example in YARN-6151.
> Since node reservations are to prevent starvation of apps requesting large
> containers, triggering these reservations only on starved applications would
> avoid this situation.
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
