[
https://issues.apache.org/jira/browse/GOBBLIN-1996?focusedWorklogId=902898&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-902898
]
ASF GitHub Bot logged work on GOBBLIN-1996:
-------------------------------------------
Author: ASF GitHub Bot
Created on: 31/Jan/24 20:38
Start Date: 31/Jan/24 20:38
Worklog Time Spent: 10m
Work Description: phet commented on code in PR #3865:
URL: https://github.com/apache/gobblin/pull/3865#discussion_r1473408282
##########
gobblin-temporal/src/main/java/org/apache/gobblin/temporal/joblauncher/GobblinJobLauncher.java:
##########
@@ -206,6 +206,13 @@ public void launchJob(@Nullable JobListener jobListener)
throws JobException {
cancelJob(jobListener);
}
} finally {
+ // NOTE: This code only makes sense when there is 1 source / workflow
being launched per application. This is a stop-gap
+ // for recreating batch job behavior in GaaS. Given the current
constraints of yarn applications requiring a static proxy user
+ // during application creation, it is not possible to have multiple
workflows running in the same application.
+ // and so it makes sense ti just kill the job after this is complete
+ log.info("Requesting the AM to shutdown after the job {} completed",
this.jobContext.getJobId());
+ eventBus.post(new ClusterManagerShutdownRequest());
Review Comment:
is it ok to always doing this, even w/o checking `isLaunched`?
##########
gobblin-temporal/src/main/java/org/apache/gobblin/temporal/joblauncher/GobblinJobLauncher.java:
##########
@@ -206,6 +206,13 @@ public void launchJob(@Nullable JobListener jobListener)
throws JobException {
cancelJob(jobListener);
}
} finally {
+ // NOTE: This code only makes sense when there is 1 source / workflow
being launched per application. This is a stop-gap
+ // for recreating batch job behavior in GaaS. Given the current
constraints of yarn applications requiring a static proxy user
+ // during application creation, it is not possible to have multiple
workflows running in the same application.
+ // and so it makes sense ti just kill the job after this is complete
+ log.info("Requesting the AM to shutdown after the job {} completed",
this.jobContext.getJobId());
+ eventBus.post(new ClusterManagerShutdownRequest());
Review Comment:
is it ok to always perform, even w/o first checking `isLaunched`?
Issue Time Tracking
-------------------
Worklog Id: (was: 902898)
Time Spent: 1h 10m (was: 1h)
> Add ability for Yarn app to terminate on finishing of temporal flow
> -------------------------------------------------------------------
>
> Key: GOBBLIN-1996
> URL: https://issues.apache.org/jira/browse/GOBBLIN-1996
> Project: Apache Gobblin
> Issue Type: Improvement
> Reporter: Matthew Ho
> Priority: Major
> Time Spent: 1h 10m
> Remaining Estimate: 0h
>
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
