[
https://issues.apache.org/jira/browse/FLINK-8942?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16406143#comment-16406143
]
ASF GitHub Bot commented on FLINK-8942:
---------------------------------------
Github user tillrohrmann commented on a diff in the pull request:
https://github.com/apache/flink/pull/5699#discussion_r175726725
--- Diff:
flink-runtime/src/main/java/org/apache/flink/runtime/heartbeat/HeartbeatManagerImpl.java
---
@@ -106,8 +107,8 @@ Executor getExecutor() {
return heartbeatListener;
}
- Collection<HeartbeatManagerImpl.HeartbeatMonitor<O>>
getHeartbeatTargets() {
- return heartbeatTargets.values();
+ Collection<Map.Entry<ResourceID, HeartbeatMonitor<O>>>
getHeartbeatTargets() {
--- End diff --
What about adding a `getTargetID()` to the `HeartbeatMonitor`? Then we
would not have to expose, even though it is an internal method, that we use a
`Map` to store the heartbeat targets. We could then keep the return type
`Collection<HeartbeatMonitor>`.
> Pass target ResourceID to HeartbeatListener#retrievePayload
> -----------------------------------------------------------
>
> Key: FLINK-8942
> URL: https://issues.apache.org/jira/browse/FLINK-8942
> Project: Flink
> Issue Type: Improvement
> Components: Distributed Coordination
> Affects Versions: 1.5.0
> Reporter: Chesnay Schepler
> Assignee: Chesnay Schepler
> Priority: Blocker
> Fix For: 1.5.0
>
>
> For FLINK-8881 we need a way to determine to which JobManager we are sending
> the heartbeats, otherwise we would have to send all accumulators, that is for
> all jobs, to each connected JobManager.
> I suggest to pass the target {{ResourceID}} to
> {{HeartbeatListener#retrievePayload}} which generates the payload.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
