[
https://issues.apache.org/jira/browse/KYLIN-3892?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16814011#comment-16814011
]
Temple Zhou commented on KYLIN-3892:
------------------------------------
You can enable this feature by setting the
kylin.job.scheduler.priority-considered=true.
When there are too many jobs running, you can submit the job with
priorityOffset property.
For example:
The number of running jobs beyonds the kylin.job.max-concurrent-jobs(the new
cubing job may be pending), meanwhile you have a cubing job with critical
importance, which is really a dilemma if we can't set the job priority.
But now, all the job will be submitted with priorityOffset 0 (default value)
and you can set the priorityOffset with positive value if the job priority is
higher than normal, on the contrary, set it with negative value.
{code:bash}
curl -X PUT \
http://localhost:7070/kylin/api/cubes/xxx/rebuild \
-H 'Authorization: Basic ***' \
-H 'Content-Type: application/json' \
-H 'cache-control: no-cache' \
-d '{
"buildType": "BUILD",
"startTime": 1552694400000,
"endTime": 1552953600000,
"priorityOffset": 10 # The higher the value, the higher the priority
}'
{code}
> Set cubing job priority
> -----------------------
>
> Key: KYLIN-3892
> URL: https://issues.apache.org/jira/browse/KYLIN-3892
> Project: Kylin
> Issue Type: New Feature
> Components: Job Engine
> Affects Versions: v2.4.0, v2.5.0, v2.6.0
> Reporter: Temple Zhou
> Assignee: Temple Zhou
> Priority: Minor
> Fix For: v2.6.2
>
>
> The cubing job with high priority will be delayed when there are too many
> tasks running.
> So I want to set the job priority for the important cubing jobs.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
