[
https://issues.apache.org/jira/browse/KYLIN-5734?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17795299#comment-17795299
]
zhong.zhu commented on KYLIN-5734:
----------------------------------
h1. Fix Design
_*org.apache.kylin.jobs.execution.ExecutableContext#addRunningJob* In this
method the logic of adding the current thread
{*}(runningJobThreads.put(executable.getId(), Thread.
currentThread()){*}{*}){*}_ independently. The purpose of this is that the
addRunningJob is just a record of which tasks have been scheduled, which is
used to determine that they should not be scheduled repeatedly, and should not
be added to the current thread (the scheduling thread), but rather when the
task is actually executed.
> Problems with task scheduling logic
> -----------------------------------
>
> Key: KYLIN-5734
> URL: https://issues.apache.org/jira/browse/KYLIN-5734
> Project: Kylin
> Issue Type: Bug
> Affects Versions: 5.0-beta
> Reporter: zhong.zhu
> Assignee: zhong.zhu
> Priority: Major
> Fix For: 5.0.0
>
>
> When a task is scheduled, the task is logged into runningJobs and the current
> thread is logged into runningJobThreads, which is expected to be the thread
> executing the task, but is actually the thread of the scheduler, which leads
> to subsequent attempts to interrupt the scheduler FetcherRunner when the task
> is killed.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
