You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@flink.apache.org by "Bill Liu (JIRA)" <ji...@apache.org> on 2017/04/13 00:40:41 UTC
[jira] [Commented] (FLINK-5668) passing taskmanager configuration
through taskManagerEnv instead of file
[ https://issues.apache.org/jira/browse/FLINK-5668?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15966934#comment-15966934 ]
Bill Liu commented on FLINK-5668:
---------------------------------
[~StephanEwen] and [~wheat9]
Sorry to raise this issue again.
I agree that Flink Runtime needs a distributed file system, but I don't get why it need to be writable.
For the example of persistent `Jobgraph`, it will be even better if user has the choice to give path of `JobGraph` directly.
Furthermore it's the Flink Client writing `Jobgraph` not the Flink Runtime.
For my case, I would like to have the `JobGraph` and jars in a http server, which are used as resources to bootstrap the Flink Job on Yarn.
So my point is the Flink Runtime shouldn't assume `yarn.deploy.fs` is writable.
> passing taskmanager configuration through taskManagerEnv instead of file
> ------------------------------------------------------------------------
>
> Key: FLINK-5668
> URL: https://issues.apache.org/jira/browse/FLINK-5668
> Project: Flink
> Issue Type: Improvement
> Components: YARN
> Reporter: Bill Liu
> Original Estimate: 48h
> Remaining Estimate: 48h
>
> When create a Flink cluster on Yarn, JobManager depends on HDFS to share taskmanager-conf.yaml with TaskManager.
> It's better to share the taskmanager-conf.yaml on JobManager Web server instead of HDFS, which could reduce the HDFS dependency at job startup.
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)