[
https://issues.apache.org/jira/browse/FLINK-8608?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16362238#comment-16362238
]
ASF GitHub Bot commented on FLINK-8608:
---------------------------------------
Github user tillrohrmann commented on a diff in the pull request:
https://github.com/apache/flink/pull/5431#discussion_r167852786
--- Diff:
flink-runtime/src/main/java/org/apache/flink/runtime/entrypoint/ClusterEntrypoint.java
---
@@ -63,6 +89,10 @@
*/
public abstract class ClusterEntrypoint implements FatalErrorHandler {
+ public static final ConfigOption<String> EXECUTION_MODE = ConfigOptions
+ .key("cluster.execution-mode")
--- End diff --
There are other configuration parameters as well like
`local.number-taskmanager`. In the old Yarn code, we used environment variables
to pass certain settings. I think the configuration is a better place. However,
I agree that this might be a bit confusing. I'll add the prefix `internal` to
the key.
> Add MiniDispatcher for job mode
> -------------------------------
>
> Key: FLINK-8608
> URL: https://issues.apache.org/jira/browse/FLINK-8608
> Project: Flink
> Issue Type: New Feature
> Components: Distributed Coordination
> Affects Versions: 1.5.0
> Reporter: Till Rohrmann
> Assignee: Till Rohrmann
> Priority: Major
> Labels: flip-6
> Fix For: 1.5.0
>
>
> In order to properly support the job mode, we need a {{MiniDispatcher}} which
> is started with a pre initialized {{JobGraph}} and launches a single
> {{JobManagerRunner}} with this job. Once the job is completed and if the
> {{MiniDispatcher}} is running in detached mode, the {{MiniDispatcher}} should
> terminate.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
