tillrohrmann commented on a change in pull request #15159:
URL: https://github.com/apache/flink/pull/15159#discussion_r597613334
##########
File path: docs/layouts/shortcodes/generated/all_jobmanager_section.html
##########
@@ -14,11 +14,17 @@
<td>Integer</td>
<td>Configure the minimum increase in parallelism for a job to
scale up.</td>
</tr>
+ <tr>
+
<td><h5>jobmanager.adaptive-scheduler.resource-stabilization-timeout</h5></td>
+ <td style="word-wrap: break-word;">10 s</td>
+ <td>Duration</td>
+ <td>The resource stabilization timeout defines the time the
JobManager will wait if fewer than the required resources are available, but
sufficient resources for execution are there. Once this timeout has passed, the
job will start executing with the available resources, or fail, if the
resources are not sufficient. The timeout starts as soon as there are
sufficient resources available for execution.<br />If <span
markdown="span">`scheduler-mode`</span> is configured to <span
markdown="span">`REACTIVE`</span>, this configuration value will default to 0,
so that jobs are starting immediately with the available resources.</td>
Review comment:
Yes, this can happen but to me the description reads a bit like we have
seen once that the resources are sufficient and then we trigger the timeout and
no matter what when the timeout triggers we try to schedule the job. Instead,
the system should still check whether it has sufficient resources and only try
to schedule the job if this is still true.
--
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:
[email protected]
