[
https://issues.apache.org/jira/browse/FLINK-9911?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16552765#comment-16552765
]
ASF GitHub Bot commented on FLINK-9911:
---------------------------------------
Github user GJL commented on a diff in the pull request:
https://github.com/apache/flink/pull/6386#discussion_r204368060
--- Diff:
flink-runtime/src/main/java/org/apache/flink/runtime/jobmaster/slotpool/SlotPool.java
---
@@ -1011,7 +1011,7 @@ public void failAllocation(final AllocationID
allocationID, final Exception caus
failPendingRequest(pendingRequest, cause);
}
else if (availableSlots.tryRemove(allocationID)) {
- log.debug("Failed available slot [{}] with ",
allocationID, cause);
+ log.debug("Failed available slot with allocation id
{}.", allocationID, cause);
--- End diff --
Why remove *allocation id*?
> SlotPool#failAllocation is called outside of main thread
> --------------------------------------------------------
>
> Key: FLINK-9911
> URL: https://issues.apache.org/jira/browse/FLINK-9911
> Project: Flink
> Issue Type: Bug
> Components: JobManager
> Affects Versions: 1.5.1, 1.6.0, 1.7.0
> Reporter: Till Rohrmann
> Assignee: Till Rohrmann
> Priority: Blocker
> Labels: pull-request-available
> Fix For: 1.5.2, 1.6.0
>
>
> The {{JobMaster}} calls directly into the {{SlotPool#failAllocation}} in the
> method {{JobMaster#notifyAllocationFailure}}. This can the {{SlotPool}} to go
> into an inconsistent state.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
