0

I am getting the below error in Spark 1.5 :

 Diagnostics: Container [pid=19554,containerID=container_e94_1518800506024_42837_02_000017] is running beyond physical memory limits. Current usage: 3.5 GB of 3.5 GB physical memory used; 4.3 GB of 7.3 GB virtual memory used. Killing container. Dump of the process-tree for container_e94_1518800506024_42837_02_000017

MASTER_URL=yarn-cluster

NUM_EXECUTORS=10

EXECUTOR_MEMORY=4G

EXECUTOR_CORES=6

DRIVER_MEMORY=3G

The data the application reads is 7MB of avro file, but there are multiple writes in spark application.

Is there any problem with Job configuration ?

NanoPish
  • 1,379
  • 1
  • 19
  • 35
Surender Raja
  • 3,553
  • 8
  • 44
  • 80
  • If possible, try increasing executors' memory. There's also probably a memory intensive action in your code which could be improved. – Xavier Guihot Mar 08 '18 at 13:14
  • Please share your code. – Vladislav Varslavans Mar 08 '18 at 13:23
  • It is big application , I wont able to share the code, It reads a 7MB input file and does some filters and joins and then stores output in 32 HDFS dirs. one by one . It stores into 32 dirs. because there are 32 partitions are there in the final dataframe . – Surender Raja Mar 08 '18 at 13:39
  • Where is your spark cluster running? Yarn? Local? Kubernetes? – TobiSH Mar 08 '18 at 13:41
  • It is yarn which is allocating resources and this spark application is running Hadoop cluster , cloudera CD5.5 (Hadoop 2.x) – Surender Raja Mar 08 '18 at 13:43
  • Possible duplicate of [Container is running beyond memory limits](https://stackoverflow.com/questions/21005643/container-is-running-beyond-memory-limits) – Xavier Guihot Mar 08 '18 at 17:57

0 Answers0