[
https://issues.apache.org/jira/browse/GOBBLIN-1728?focusedWorklogId=818927&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-818927
]
ASF GitHub Bot logged work on GOBBLIN-1728:
-------------------------------------------
Author: ASF GitHub Bot
Created on: 20/Oct/22 19:52
Start Date: 20/Oct/22 19:52
Worklog Time Spent: 10m
Work Description: homatthew commented on code in PR #3586:
URL: https://github.com/apache/gobblin/pull/3586#discussion_r1001052159
##########
gobblin-yarn/src/main/java/org/apache/gobblin/yarn/YarnService.java:
##########
@@ -845,8 +877,6 @@ public void onContainersAllocated(List<Container>
containers) {
instanceName = null;
}
}
- allocatedContainerCountMap.put(containerHelixTag,
Review Comment:
I re-read the API and I think you're correct.
https://hadoop.apache.org/docs/stable/hadoop-yarn/hadoop-yarn-client/apidocs/org/apache/hadoop/yarn/client/api/async/AMRMClientAsync.CallbackHandler.html
Because this method is triggered by heartbeat and this method is called
after the `onContainersCompleted`, maybe any implementation would work here
since it's not even concurrent lol. All writes to this map actually seem single
threaded
Issue Time Tracking
-------------------
Worklog Id: (was: 818927)
Time Spent: 3h 10m (was: 3h)
> Yarn Service requests too many containers due to improper calculation
> ---------------------------------------------------------------------
>
> Key: GOBBLIN-1728
> URL: https://issues.apache.org/jira/browse/GOBBLIN-1728
> Project: Apache Gobblin
> Issue Type: New Feature
> Reporter: Matthew Ho
> Priority: Major
> Time Spent: 3h 10m
> Remaining Estimate: 0h
>
> Yarn Service is responsible for calculating the number of instances based on
> the helix tasks. Yarn service tracks the number of instances by asking Yarn
> for the number of resource requests and the number of allocated containers.
>
> It uses this count to determine if it should ask for more containers or
> shrink the number of containers. This calculation is currently done
> improperly and we continue to request containers when we have enough
> requested.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
