Skip to content

Latest commit

 

History

History
128 lines (77 loc) · 5.97 KB

troubleshooting.md

File metadata and controls

128 lines (77 loc) · 5.97 KB
Raw

Troubleshooting

Tests don't pass or have timeouts

If you used reStart to start a local job server, be sure it's stopped using reStop first.

Also, try adding -Dakka.test.timefactor=X to SBT_OPTS before launching sbt, where X is a number greater than 1. This scales out the Akka TestKit timeouts by a factor X.

Requests are timing out

or you get akka.pattern.AskTimeoutException.

send timeout param along with your request (in secs). eg below.

http://devsparkcluster.cloudapp.net/jobs?appName=job-server-tests&classPath=spark.jobserver.WordCountExample&sync=true&timeout=20

You may need to adjust Spray's default request timeout and idle timeout, which are by default 40 secs and 60 secs. To do this, modify the configuration file in your deployed job server, adding a section like the following:

spray.can.server {
  idle-timeout = 210 s
  request-timeout = 200 s
}

Then simply restart the job server.

Note that the idle-timeout must be higher than request-timeout, or Spray and the job server won't start.

Timeout getting large job results

If your job returns a large job result, it may exceed Akka's maximum network message frame size, in which case the result is dropped and you may get a network timeout. Change the following configuration, which defaults to 10 MiB:

akka.remote.netty.tcp.maximum-frame-size = 100 MiB

Job with status finished has no result

On jobs with large results or many concurrent jobs, the REST API at /job/abc.. might return status FINISHED but does not contain any result. This might happen in three cases:

  1. The job finished right now and results are in transfer between the akka actors. Try it again some milliseconds later.
  2. The job finished some time ago and results are removed from results cache already. See spark.jobserver.job-result-cache-size to increase the cache size.
  3. The spark jobserver instance restarted which clears the results cache.

AskTimeout when starting job server or contexts

If you are loading large jars or dependent jars, either at startup or when creating a large context, the database such as H2 may take a really long time to write those bytes to disk. You need to adjust the context timeout setting:

spark.jobserver.context-creation-timeout

Set it to 60 seconds or longer, especially if your jars are in the many MBs.

NOTE: if you are running SJS in Docker, esp on AWS, you might need to enable host-only networking.

Job server won't start / cannot bind to 0.0.0.0:8090

Check that another process isn't already using that port. If it is, you may want to start it on another port:

reStart --- -Dspark.jobserver.port=2020

Job Server Doesn't Connect to Spark Cluster

Finally, I got the problem solved. There are two problems in my configuration:

  1. the version of spark cluster is 1.1 but the spark version in job server machine is 1.0.2 after upgrading spark to 1.1 in job server machine, jobs can be submitted to spark cluster (can show in spark UI) but cannot be executed.
  2. the spark machines need to know the host name of job server machine after this fixed, I can run jobs submitted from a remote job server successfully.

(Thanks to @pcliu)

Exception in thread "main" java.lang.NoSuchMethodError: akka.actor.ActorRefFactory.dispatcher()Lscala/concurrent/ExecutionContextExecutor;

If you are running CDH 5.3 or older, you may have an incompatible version of Akka bundled together. :( Fortunately, one of our users has put together a branch that works ... try that out!

(Older instructions) Try modifying the version of Akka included with spark-jobserver to match the one in CDH (2.2.4, I think), or upgrade to CDH 5.4. If you are on CDH 5.4, check that sparkVersion in Dependencies.scala matches CDH. Or see isse #154.

java.lang.NoSuchMethodError: org.joda.time.DateTime.now()

This time the problem is caused by incompatible class versions of the joda.time package in Hive and the Spark Job Server on Cloudera (java.lang.NoSuchMethodError: org.joda.time.DateTime.now()Lorg/joda/time/DateTime exception in the spark job server log). To solve the problem execute the following two commands on the machine the Job Server is installed:

sed -i -e 's#--driver-class-path.*SPARK_HOME/../hive/lib/.*##' /opt/spark-job-server/manager_start.sh
sed -i -e 's#--driver-class-path.*SPARK_HOME/../hive/lib/.*##' /opt/spark-job-server/server_start.sh

This removes the problematic driver class path entries from the two spark job server scripts. (from @koetter)

I am running CDH 5.3 and Job Server doesn't work

See above.

I want to run job-server on Windows

  1. Create directory C:\Hadoop\bin
  2. Download http://public-repo-1.hortonworks.com/hdp-win-alpha/winutils.exe and place it in C:\Hadoop\bin
  3. Set environment variable HADOOP_HOME (either in a .bat script or within OS properties) HADOOP_HOME=C:\Hadoop
  4. Start spark-job-server in a shell that has the HADOOP_HOME environment set.
  5. Submit the WordCountExample Job.

(Thanks to Javier Delgadillo)

Akka Deadletters / Workers disconnect from Job Server

Most likely a networking issue. Try using IP addresses instead of DNS. (happens in AWS)

java.lang.ClassNotFoundException when staring spark-jobserver from sbt

Symptom:

You start from SBT using reStart, and when try to create a HiveContext or SQLContext, e.g. using context-factory=spark.jobserver.context.HiveContextFactory and get an error like this {"status": "CONTEXT INIT ERROR", "result": { "message": "spark.jobserver.context.HiveContextFactory", "errorClass": "java.lang.ClassNotFoundException", ... } ...

Solution:

Before typing reStart in sbt, type project job-server-extras and only then start it using reStart

Accessing a config file in my job jar

ConfigFactory.parseReader(paramReader = new InputStreamReader(getClass().getResourceAsStream(s"/$myPassedConfigPath"))