[GitHub] [zeppelin] zjffdu commented on a change in pull request #4255: [ZEPPELIN-5560] spark yarn app end with failed status in yarn-cluster mode

[GitBox <gi...@apache.org>]

zjffdu commented on a change in pull request #4255:
URL: https://github.com/apache/zeppelin/pull/4255#discussion_r730682888



##########
File path: zeppelin-zengine/src/main/java/org/apache/zeppelin/interpreter/launcher/SparkInterpreterLauncher.java
##########
@@ -98,9 +98,14 @@ public SparkInterpreterLauncher(ZeppelinConfiguration zConf, RecoveryStorage rec
       sparkProperties.setProperty("spark.pyspark.python", condaEnvName + "/bin/python");
     }
 
-    if (isYarnMode() && getDeployMode().equals("cluster")) {
+    if (isYarnCluster()) {
       env.put("ZEPPELIN_SPARK_YARN_CLUSTER", "true");
       sparkProperties.setProperty("spark.yarn.submit.waitAppCompletion", "false");
+      // Need to set `zeppelin.interpreter.forceShutdown` in interpreter properties directly
+      // instead of updating sparkProperties.
+      // Because `zeppelin.interpreter.forceShutdown` is initialized in RemoteInterpreterServer
+      // before SparkInterpreter is created.
+      context.getProperties().put("zeppelin.interpreter.forceShutdown", "false");

Review comment:
       It would pass to RemoteInterpreterServer when creating interpreter.
   https://github.com/apache/zeppelin/blob/master/zeppelin-interpreter/src/main/java/org/apache/zeppelin/interpreter/remote/RemoteInterpreterServer.java#L356




-- 
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: dev-unsubscribe@zeppelin.apache.org

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