Page History

...

Pig on Tez

Cascading on Tez

TBD

Enable logging for specific tasks in Tez

• "tez.task-specific.launch.cmd-opts.list" and "tez.task-specific.log.level" can be used to control log levels for specific tasks in the vertex of a DAG. Examples are given below
  
  • tez.task-specific.launch.cmd-opts.list="Map 1[0]" tez.task-specific.log.level="INFO;org.apache=DEBUG;"
    • Enable DEBUG log for "org.apache" for task-0 in Map 1.  Rest of the logs would be at INFO level

Profiling in Tez

It is possible to profile specific tasks in specific vertices in Tez. Examples are given below.

Example A: Use Xprof. Add "-Xprof" to tez.task.launch.cmd-opts, e.g.  tez.task.launch.cmd-opts="-Xprof".

Example B. Use yourkit on specific tasks via "tez.task-specific.launch.cmd-opts.list" and "tez.task-specific.launch.cmd-opts". If your cluster doesn't have yourkit installed, you can config "tez.lib.uris" and "tez.lib.uris.classpath" so that the yourkit tar.gz is localized by yarn and added to the container classpath. With the below configuration, yourkit snapshot files will be created under yarn container log directory based on vertex name and task index. Then you can use "yarn logs" command to download the container log and exact the snapshot portion from the log.

• tez.task-specific.launch.cmd-opts.list
  • Specifies the tasks in different vertices where additional options have to be specified
  • Examples:
    • tez.task-specific.launch.cmd-opts.list=“M5[0]" - Specifies task 0 in vertex M5
      
    • tez.task-specific.launch.cmd-opts.list=“Map10[5,20]" - Specifies task 5,20 in vertex Map 10
    • tez.task-specific.launch.cmd-opts.list=“M5[]" - Specifies all tasks in M5
    • tez.task-specific.launch.cmd-opts.list=“M5[1:3,10]" - Specifies 1,2,3,10 tasks in M5
      • However, defining partial ranges like "M5[:3]" is not yet supported.
    • tez.task-specific.launch.cmd-opts.list=“M5[0];V2[10]" - Specifies task 0 in vertex M5 and task 10 in vertex V2
      
      
• tez.task-specific.launch.cmd-opts
  • Specifies the additional task specific JVM launch options that need to be added.
  • __VERTEX_NAME__ and __TASK_INDEX__ can be specified in the option, which would be replaced at runtime.
  • Examples:
    • tez.task-specific.launch.cmd-opts="-agentpath:yjp-2016.02-b46-linux.tar.gz/yjp-2016.02/bin/linux-x86-64/libyjpagent.so=disablej2ee,sampling,disabletracing,disablealloc,periodicperf=60,dir=<LOG_DIR>/__VERTEX_NAME__/__TASK_INDEX__"

• tez.lib.uris

  • Add the yourkit tar.gz HDFS path to tez.lib.uris. Example: tez.lib.uris=/lib/tez-0.9.0-SNAPSHOT-minimal.tar.gz,/lib/yjp-2016.02-b46-linux.tar.gz

• tez.lib.uris.classpath

  • tez runtime tar.gz is defined in tez.lib.uris, config the classpath for the tez runtime tar.gz via tez.lib.uris.classpath=tez-0.9.0-SNAPSHOT-minimal.tar.gz/*,tez-0.9.0-SNAPSHOT-minimal.tar.gz/lib/.