[jira] [Commented] (LIVY-541) Multiple Livy servers submitting to Yarn results in LivyException: Session is finished ... No YARN application is found with tag livy-session-197-uveqmqyj in 300 seconds. Please check your cluster status, it is may be very busy

["Jeffrey(Xilang) Yan (Jira)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/LIVY-541?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16953290#comment-16953290 ] 

Jeffrey(Xilang) Yan commented on LIVY-541:
------------------------------------------

If I remember correctly, Livy looking for app on yarn with tag, not application ID, which is show in log :No YARN application is found with *tag* livy-session-197-uveqmqyj in 300 seconds. 
Could you reproduce with 0.6 release, and have you checked if your application on yarn is started?

> Multiple Livy servers submitting to Yarn results in LivyException: Session is finished ... No YARN application is found with tag livy-session-197-uveqmqyj in 300 seconds. Please check your cluster status, it is may be very busy
> -----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
>
>                 Key: LIVY-541
>                 URL: https://issues.apache.org/jira/browse/LIVY-541
>             Project: Livy
>          Issue Type: Bug
>          Components: Server
>    Affects Versions: 0.5.0
>         Environment: Hortonworks HDP 2.6
>            Reporter: Hari Sekhon
>            Priority: Critical
>          Time Spent: 10m
>  Remaining Estimate: 0h
>
> It appears Livy doesn't differentiate sessions properly in Yarn causing errors when running multiple Livy servers behind a load balancer for HA / performance scaling on the same Hadoop cluster.
> Each livy server uses monotonically incrementing session IDs with a random suffix but it appears that the random suffix isn't passed to Yarn which results in the following errors on the Livy server which is further behind in session numbers because it appears to find the session with the same number has already finished (submitted earlier by a different user on another Livy server as seen in Yarn RM UI):
> {code:java}
> org.apache.zeppelin.livy.LivyException: Session 197 is finished, appId: null, log: [	at org.apache.hadoop.fs.FileSystem$Cache.closeAll(FileSystem.java:2887), at org.apache.hadoop.fs.FileSystem$Cache$ClientFinalizer.run(FileSystem.java:2904), at java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:511), at java.util.concurrent.FutureTask.run(FutureTask.java:266), at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1142), at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:617), at java.lang.Thread.run(Thread.java:748), 
> YARN Diagnostics: , java.lang.Exception: No YARN application is found with tag livy-session-197-uveqmqyj in 300 seconds. Please check your cluster status, it is may be very busy., org.apache.livy.utils.SparkYarnApp.org$apache$livy$utils$SparkYarnApp$$getAppIdFromTag(SparkYarnApp.scala:182) org.apache.livy.utils.SparkYarnApp$$anonfun$1$$anonfun$4.apply(SparkYarnApp.scala:239) org.apache.livy.utils.SparkYarnApp$$anonfun$1$$anonfun$4.apply(SparkYarnApp.scala:236) scala.Option.getOrElse(Option.scala:120) org.apache.livy.utils.SparkYarnApp$$anonfun$1.apply$mcV$sp(SparkYarnApp.scala:236) org.apache.livy.Utils$$anon$1.run(Utils.scala:94)]
> at org.apache.zeppelin.livy.BaseLivyInterpreter.createSession(BaseLivyInterpreter.java:300)
> at org.apache.zeppelin.livy.BaseLivyInterpreter.initLivySession(BaseLivyInterpreter.java:184)
> at org.apache.zeppelin.livy.LivySharedInterpreter.open(LivySharedInterpreter.java:57)
> at org.apache.zeppelin.interpreter.LazyOpenInterpreter.open(LazyOpenInterpreter.java:69)
> at org.apache.zeppelin.livy.BaseLivyInterpreter.getLivySharedInterpreter(BaseLivyInterpreter.java:165)
> at org.apache.zeppelin.livy.BaseLivyInterpreter.open(BaseLivyInterpreter.java:139)
> at org.apache.zeppelin.interpreter.LazyOpenInterpreter.open(LazyOpenInterpreter.java:69)
> at org.apache.zeppelin.interpreter.remote.RemoteInterpreterServer$InterpretJob.jobRun(RemoteInterpreterServer.java:493)
> at org.apache.zeppelin.scheduler.Job.run(Job.java:175)
> at org.apache.zeppelin.scheduler.FIFOScheduler$1.run(FIFOScheduler.java:139)
> at java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:511)
> at java.util.concurrent.FutureTask.run(FutureTask.java:266)
> at java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$201(ScheduledThreadPoolExecutor.java:180)
> at java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:293)
> at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1142)
> at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:617)
> at java.lang.Thread.run(Thread.java:748)
> {code}



--
This message was sent by Atlassian Jira
(v8.3.4#803005)