[ https://issues.apache.org/jira/browse/SPARK-12263?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15051029#comment-15051029 ]
Sean Owen commented on SPARK-12263: ----------------------------------- The message already gives you the solution though. Yes, I'm also saying that stating the minimum (which is effectively 1MB) is a good idea in that same message. It would be more work to record where the setting came from and report the original raw value; I don't think it's worth it for this case. > IllegalStateException: Memory can't be 0 for SPARK_WORKER_MEMORY without unit > ----------------------------------------------------------------------------- > > Key: SPARK-12263 > URL: https://issues.apache.org/jira/browse/SPARK-12263 > Project: Spark > Issue Type: Improvement > Components: Documentation > Reporter: Jacek Laskowski > Priority: Trivial > > When starting a worker with the following command - note > {{SPARK_WORKER_MEMORY=1024}} it fails saying that the memory was 0 while it > was 1024 (without size unit). > {code} > ➜ spark git:(master) ✗ SPARK_WORKER_MEMORY=1024 SPARK_WORKER_CORES=5 > ./sbin/start-slave.sh spark://localhost:7077 > starting org.apache.spark.deploy.worker.Worker, logging to > /Users/jacek/dev/oss/spark/logs/spark-jacek-org.apache.spark.deploy.worker.Worker-1-japila.local.out > failed to launch org.apache.spark.deploy.worker.Worker: > INFO ShutdownHookManager: Shutdown hook called > INFO ShutdownHookManager: Deleting directory > /private/var/folders/0w/kb0d3rqn4zb9fcc91pxhgn8w0000gn/T/spark-f4e5f222-e938-46b2-a189-241453cf1f50 > full log in > /Users/jacek/dev/oss/spark/logs/spark-jacek-org.apache.spark.deploy.worker.Worker-1-japila.local.out > {code} > The full stack trace is as follows: > {code} > Setting default log level to "WARN". > To adjust logging level use sc.setLogLevel(newLevel). > INFO Worker: Registered signal handlers for [TERM, HUP, INT] > Exception in thread "main" java.lang.IllegalStateException: Memory can't be > 0, missing a M or G on the end of the memory specification? > at > org.apache.spark.deploy.worker.WorkerArguments.checkWorkerMemory(WorkerArguments.scala:179) > at > org.apache.spark.deploy.worker.WorkerArguments.<init>(WorkerArguments.scala:64) > at org.apache.spark.deploy.worker.Worker$.main(Worker.scala:691) > at org.apache.spark.deploy.worker.Worker.main(Worker.scala) > INFO ShutdownHookManager: Shutdown hook called > INFO ShutdownHookManager: Deleting directory > /private/var/folders/0w/kb0d3rqn4zb9fcc91pxhgn8w0000gn/T/spark-f4e5f222-e938-46b2-a189-241453cf1f50 > {code} > The following command starts spark standalone worker successfully: > {code} > SPARK_WORKER_MEMORY=1g SPARK_WORKER_CORES=5 ./sbin/start-slave.sh > spark://localhost:7077 > {code} > The master reports: > {code} > INFO Master: Registering worker 192.168.1.6:63884 with 5 cores, 1024.0 MB RAM > {code} -- This message was sent by Atlassian JIRA (v6.3.4#6332) --------------------------------------------------------------------- To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org