Is there any difference between executor-cores
and spark.executor.cores
used in spark-submit command?
My job is failed because of GC overhead memory error, so i am trying to increase cores and memory settings.
The total volume i am processing is 50 M records in two files.
The executor-cores
flag used in the spark-submit
command simply set the spark.executor.cores
setting on the Spark-configuration. So they have the same effect :)
A couple of things you may try:
1) You have tagged the question with YARN, so if you find that you are not utilizing all of your cores, you should have a look at Apache Hadoop Yarn - Underutilization of cores
2) Many memory issues on YARN are solved when you increase the memory-overhead by explicitly setting spark.yarn.executor.memoryOverhead. It will default to max(386MB, 0.10* executorMemory)
and that is frequently not enough.
The technical post webpages of this site follow the CC BY-SA 4.0 protocol. If you need to reprint, please indicate the site URL or the original address.Any question please contact:yoyou2525@163.com.