You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@celeborn.apache.org by an...@apache.org on 2023/04/27 06:16:35 UTC

[incubator-celeborn] branch main updated: [CELEBORN-560][BUG] Rerun task in spark later then RSS stageEnd cause NPE then job failed (#1466)

This is an automated email from the ASF dual-hosted git repository.

angerszhuuuu pushed a commit to branch main
in repository https://gitbox.apache.org/repos/asf/incubator-celeborn.git


The following commit(s) were added to refs/heads/main by this push:
     new ce21a738a [CELEBORN-560][BUG] Rerun task in spark later then RSS stageEnd cause NPE then job failed (#1466)
ce21a738a is described below

commit ce21a738a92e0fdadddf7194cea4d9020d7c1250
Author: Angerszhuuuu <an...@gmail.com>
AuthorDate: Thu Apr 27 14:16:32 2023 +0800

    [CELEBORN-560][BUG] Rerun task in spark later then RSS stageEnd cause NPE then job failed (#1466)
---
 .../src/main/scala/org/apache/celeborn/client/LifecycleManager.scala  | 4 ----
 1 file changed, 4 deletions(-)

diff --git a/client/src/main/scala/org/apache/celeborn/client/LifecycleManager.scala b/client/src/main/scala/org/apache/celeborn/client/LifecycleManager.scala
index a82ea48f8..54d7ed07b 100644
--- a/client/src/main/scala/org/apache/celeborn/client/LifecycleManager.scala
+++ b/client/src/main/scala/org/apache/celeborn/client/LifecycleManager.scala
@@ -569,9 +569,6 @@ class LifecycleManager(appId: String, val conf: CelebornConf) extends RpcEndpoin
     }
 
     if (commitManager.tryFinalCommit(shuffleId)) {
-      // release resources and clear worker info
-      shuffleAllocatedWorkers.remove(shuffleId)
-
       requestReleaseSlots(
         rssHARetryClient,
         ReleaseSlots(applicationId, shuffleId, List.empty.asJava, List.empty.asJava))
@@ -631,7 +628,6 @@ class LifecycleManager(appId: String, val conf: CelebornConf) extends RpcEndpoin
     if (shuffleResourceExists(shuffleId)) {
       logWarning(s"Partition exists for shuffle $shuffleId, " +
         "maybe caused by task rerun or speculative.")
-      shuffleAllocatedWorkers.remove(shuffleId)
       requestReleaseSlots(
         rssHARetryClient,
         ReleaseSlots(appId, shuffleId, List.empty.asJava, List.empty.asJava))