[GitHub] [hadoop] goiri commented on a diff in pull request #5672: YARN-7720. Race condition between second app attempt and UAM timeout when first attempt node is down.

["goiri (via GitHub)" <gi...@apache.org>]

goiri commented on code in PR #5672:
URL: https://github.com/apache/hadoop/pull/5672#discussion_r1198012980


##########
hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/ResourceManager.java:
##########
@@ -703,6 +703,24 @@ protected static void validateConfigs(Configuration conf) {
           + ", " + YarnConfiguration.RM_NM_HEARTBEAT_INTERVAL_MS + "="
           + heartbeatIntvl);
     }
+
+    if (HAUtil.isFederationEnabled(conf)) {
+      /*
+       * In Yarn Federation, we need UAMs in secondary sub-clusters to stay
+       * alive when the next attempt AM in home sub-cluster gets launched. If
+       * the previous AM died because the node is lost after NM timeout. It will
+       * already be too late if AM timeout is even shorter.
+       */
+      int amExpireIntvl = conf.getInt(YarnConfiguration.RM_AM_EXPIRY_INTERVAL_MS,

Review Comment:
   Can we switch to getTimeDuration() or would that break backwards compatibility.



-- 
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: common-issues-unsubscribe@hadoop.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: common-issues-unsubscribe@hadoop.apache.org
For additional commands, e-mail: common-issues-help@hadoop.apache.org