[
https://issues.apache.org/jira/browse/GOBBLIN-1728?focusedWorklogId=818933&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-818933
]
ASF GitHub Bot logged work on GOBBLIN-1728:
-------------------------------------------
Author: ASF GitHub Bot
Created on: 20/Oct/22 20:13
Start Date: 20/Oct/22 20:13
Worklog Time Spent: 10m
Work Description: homatthew commented on code in PR #3586:
URL: https://github.com/apache/gobblin/pull/3586#discussion_r1001069688
##########
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:
But then again the comment above says
```
Ensure that updates to unusedHelixInstanceNames are visible to other threads
that might concurrently
invoke the callback on container allocation.
```
Which implies there could be multiple callers to this single object. In that
case, why don't we have a lock for oncomplete method which decrements this
count?
Either way, the code is so painful to reason about when we have all these
concurrent maps + explicit locks... Gives a sense of "safety" but still is very
error prone IMO
Issue Time Tracking
-------------------
Worklog Id: (was: 818933)
Time Spent: 3.5h (was: 3h 20m)
> 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: 3.5h
> 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)
