s0nskar commented on code in PR #2711:
URL: https://github.com/apache/celeborn/pull/2711#discussion_r1774787465
##########
worker/src/main/scala/org/apache/celeborn/service/deploy/worker/Worker.scala:
##########
@@ -435,6 +435,14 @@ private[celeborn] class Worker(
0
}
}
+ // Unreleased shuffle count when worker was decommissioned
+ workerSource.addGauge(WorkerSource.UNRELEASED_SHUFFLE_COUNT) { () =>
+ if (shutdown.get() && workerStatusManager.currentWorkerStatus.getState ==
State.Exit) {
Review Comment:
After transitioning to `InDecommission` state it will wait for
`workerDecommissionForceExitTimeout` to cleanup the shuffle. My thought is that
we should only check the count of shuffle which was not able to get released
till the timeout.
Suggested condition will mean "Unreleased shuffle count when worker started
decommissioning".
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
