[
https://issues.apache.org/jira/browse/YARN-6451?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15966500#comment-15966500
]
Chris Douglas commented on YARN-6451:
-------------------------------------
Cool, I hadn't seen the {{javax.script}} package before. Throwing a bespoke
exception can also be configured to halt the JVM and call back to a debugger,
which is a nice touch for the SLS case.
* The invariants can be precompiled, to avoid the parsing/compilation overhead
for each iteration.
* If not invoking a debugger, then it'd be nice to know the bindings when the
invariant doesn't hold.
* The invariant check could be part of the {{metrics2.MetricsCollector}},
particularly if it's possible to filter the metrics it gathers based on the
configured invariants.
> Create a monitor to check whether we maintain RM (scheduling) invariants
> ------------------------------------------------------------------------
>
> Key: YARN-6451
> URL: https://issues.apache.org/jira/browse/YARN-6451
> Project: Hadoop YARN
> Issue Type: Bug
> Reporter: Carlo Curino
> Assignee: Carlo Curino
> Attachments: YARN-6451.v0.patch, YARN-6451.v1.patch,
> YARN-6451.v2.patch
>
>
> For SLS runs, as well as for live test clusters (and maybe prod), it would be
> useful to have a mechanism to continuously check whether core invariants of
> the RM/Scheduler are respected (e.g., no priority inversions, fairness mostly
> respected, certain latencies within expected range, etc..)
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]