[jira] [Created] (CARBONDATA-1366) When sort_scope=global_sort, use 'StorageLevel.MEMORY_AND_DISK_SER' instead of 'StorageLevel.MEMORY_AND_DISK' for 'convertRDD' persisting to improve loading performance

[Akash R Nilugal (Jira)]

Zhichao  Zhang created CARBONDATA-1366:
------------------------------------------

             Summary: When sort_scope=global_sort, use 'StorageLevel.MEMORY_AND_DISK_SER' instead of 'StorageLevel.MEMORY_AND_DISK' for 'convertRDD' persisting  to improve loading performance
                 Key: CARBONDATA-1366
                 URL: https://issues.apache.org/jira/browse/CARBONDATA-1366
             Project: CarbonData
          Issue Type: Bug
          Components: data-load, spark-integration
    Affects Versions: 1.2.0
            Reporter: Zhichao  Zhang
            Assignee: Zhichao  Zhang
            Priority: Minor
             Fix For: 1.2.0


My testing env and configs are as followings:

Env:
6 executors, 9G mem + 6 cores per executor

Configs:
SINGLE_PASS=true
SORT_SCOPE=GLOBAL_SORT
spark.memory.fraction=0.5

if using 'convertRDD.persist(StorageLevel.MEMORY_AND_DISK_SER)' in method 'org.apache.carbondata.spark.load.DataLoadProcessBuilderOnSpark.loadDataUsingGlobalSort', it takes about 7.2 min to load 144136697 lines (10.9 G parquet files), and if using 'convertRDD.persist(StorageLevel.MEMORY_AND_DISK)', it takes about 9.5 min to load 144136697 lines.





--
This message was sent by Atlassian JIRA
(v6.4.14#64029)