YARN-2992. ZKRMStateStore crashes due to session expiry. Contributed by Karthik Kambatla
Project: http://git-wip-us.apache.org/repos/asf/hadoop/repo Commit: http://git-wip-us.apache.org/repos/asf/hadoop/commit/71dc3449 Tree: http://git-wip-us.apache.org/repos/asf/hadoop/tree/71dc3449 Diff: http://git-wip-us.apache.org/repos/asf/hadoop/diff/71dc3449 Branch: refs/heads/HDFS-EC Commit: 71dc34499910b651a40bde1ed9a68e91819327c3 Parents: 3082f4f Author: Jian He <[email protected]> Authored: Fri Dec 26 17:59:58 2014 -0800 Committer: Zhe Zhang <[email protected]> Committed: Mon Dec 29 10:34:38 2014 -0800 ---------------------------------------------------------------------- hadoop-yarn-project/CHANGES.txt | 3 +++ .../yarn/server/resourcemanager/recovery/ZKRMStateStore.java | 3 +++ 2 files changed, 6 insertions(+) ---------------------------------------------------------------------- http://git-wip-us.apache.org/repos/asf/hadoop/blob/71dc3449/hadoop-yarn-project/CHANGES.txt ---------------------------------------------------------------------- diff --git a/hadoop-yarn-project/CHANGES.txt b/hadoop-yarn-project/CHANGES.txt index 123ac2f..9fb94d2 100644 --- a/hadoop-yarn-project/CHANGES.txt +++ b/hadoop-yarn-project/CHANGES.txt @@ -289,6 +289,9 @@ Release 2.7.0 - UNRELEASED YARN-2988. Graph#save() may leak file descriptors. (Ted Yu via ozawa) + YARN-2992. ZKRMStateStore crashes due to session expiry. (Karthik Kambatla + via jianhe) + Release 2.6.0 - 2014-11-18 INCOMPATIBLE CHANGES http://git-wip-us.apache.org/repos/asf/hadoop/blob/71dc3449/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/recovery/ZKRMStateStore.java ---------------------------------------------------------------------- diff --git a/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/recovery/ZKRMStateStore.java b/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/recovery/ZKRMStateStore.java index e559e4c..285f3d0 100644 --- a/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/recovery/ZKRMStateStore.java +++ b/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/recovery/ZKRMStateStore.java @@ -1056,6 +1056,8 @@ public class ZKRMStateStore extends RMStateStore { switch (code) { case CONNECTIONLOSS: case OPERATIONTIMEOUT: + case SESSIONEXPIRED: + case SESSIONMOVED: return true; default: break; @@ -1084,6 +1086,7 @@ public class ZKRMStateStore extends RMStateStore { if (shouldRetry(ke.code()) && ++retry < numRetries) { LOG.info("Retrying operation on ZK. Retry no. " + retry); Thread.sleep(zkRetryInterval); + createConnection(); continue; } LOG.info("Maxed out ZK retries. Giving up!");
