[
https://issues.apache.org/jira/browse/YARN-6648?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16336478#comment-16336478
]
Botong Huang commented on YARN-6648:
------------------------------------
Thanks [~curino] for the feedback with lots of good points! I've fixed them in
v4 patch, and some explanations below.
* 3.2. Changed running interval default to -1 meaning to disable the service.
Yes, GPG will probably need some admin API to start/disable services. I prefer
to leave it to future patches.
* 4.1. This is some legacy code when we initially ran into the time zone
problem. The StateStore script in trunk now already have the UTC fixes, so this
extra check is not needed any more.
* 4.3. Right now if the sub-cluster state is already unusable, we leave it
unchanged there. We only mark timed out usable ones as LOST (unusable) for now.
* 5. I decided to remove this code entirely. It is more straightforward to use
one ScheduledExecutorService to schedule all runnable tasks. In terms of
exception from the runnable, I think it can handle it properly without
crashing. Also, the SubClusterCleaner.run() also tries to never throw anything
out.
> [GPG] Add SubClusterCleaner in Global Policy Generator
> ------------------------------------------------------
>
> Key: YARN-6648
> URL: https://issues.apache.org/jira/browse/YARN-6648
> Project: Hadoop YARN
> Issue Type: Sub-task
> Reporter: Botong Huang
> Assignee: Botong Huang
> Priority: Minor
> Labels: federation, gpg
> Attachments: YARN-6648-YARN-2915.v1.patch,
> YARN-6648-YARN-7402.v2.patch, YARN-6648-YARN-7402.v3.patch
>
>
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
