[ 
https://issues.apache.org/jira/browse/KYLIN-5739?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17795175#comment-17795175
 ] 

zhong.zhu commented on KYLIN-5739:
----------------------------------

h1. Dev Degisn

First of all, it should be noted that *_kylin.jobs.max-concurrent-jobs_* can be 
set to 0, with a range of values >= 0, with real use cases.

For example: when starting Kylin, not 0, created a batch of tasks, at this 
time, I want to jam the task to prevent it from being adjusted up, so the 
project-level configuration is set to 0, then if you reboot later, init 
scheduler when the core size is 0.
h2. Root Cause

When +_{{NDefaultScheduler}}_+ is initialized for the first time, 
{*}_kylin.job.max-concurrent-jobs_{*}=0 is configured, causing an invalid 
parameter exception to be thrown during jobPool initialization.

This means that the following context is also null, and the fetcherPool's timed 
scheduling task won't get up properly.

So when you subsequently want to schedule tasks via NDefaultScheduler, since 
the internal jobPool and context are both null, you'll get NEP exceptions where 
you use the context and thread pool.

 
h3. Problematic phenomena:

After startup, all the jobs are pending, even if you change the project level 
configuration max job > 0, it doesn't work, because the thread pool is all null.

Clicking restart will report a null pointer because the context is null.

> The job concurrency number is set to 0, resulting in initialization failure, 
> and clicking the restart job button will report an error
> -------------------------------------------------------------------------------------------------------------------------------------
>
>                 Key: KYLIN-5739
>                 URL: https://issues.apache.org/jira/browse/KYLIN-5739
>             Project: Kylin
>          Issue Type: Bug
>    Affects Versions: 5.0-beta
>            Reporter: zhong.zhu
>            Assignee: zhong.zhu
>            Priority: Major
>             Fix For: 5.0.0
>
>
> Only one Kylin node was started and clicking the restart task button reported 
> an error



--
This message was sent by Atlassian Jira
(v8.20.10#820010)

Reply via email to