You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@spark.apache.org by "hantiantian (JIRA)" <ji...@apache.org> on 2019/02/01 01:22:00 UTC
[jira] [Updated] (SPARK-26726) Synchronize the amount of memory
used by the broadcast variable to the UI display
[ https://issues.apache.org/jira/browse/SPARK-26726?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
hantiantian updated SPARK-26726:
--------------------------------
Description:
The amount of memory used by the broadcast variable is not synchronized to the UI display，
spark-sql> select /*+ broadcast(a)*/ a.id,b.id from a join b on a.id = b.id;
View the app's driver log：
2019-01-25 16:45:23,726 INFO org.apache.spark.storage.BlockManagerInfo: Added broadcast_4_piece0 in memory on 10.43.xx.xx:33907 (size: 6.6 KB, free: 2.5 GB)
2019-01-25 16:45:23,727 INFO org.apache.spark.storage.BlockManagerInfo: Added broadcast_4_piece0 in memory on 10.43.xx.xx:38399 (size: 6.6 KB, free: 2.5 GB)
2019-01-25 16:45:23,745 INFO org.apache.spark.storage.BlockManagerInfo: Added broadcast_3_piece0 in memory on 10.43.xx.xx:33907 (size: 32.1 KB, free: 2.5 GB)
2019-01-25 16:45:23,749 INFO org.apache.spark.storage.BlockManagerInfo: Added broadcast_3_piece0 in memory on 10.43.xx.xx:38399 (size: 32.1 KB, free: 2.5 GB)
2019-01-25 16:45:23,838 INFO org.apache.spark.storage.BlockManagerInfo: Added broadcast_2_piece0 in memory on 10.43.xx.xx:38399 (size: 147.0 B, free: 2.5 GB)
2019-01-25 16:45:23,840 INFO org.apache.spark.storage.BlockManagerInfo: Added broadcast_2_piece0 in memory on 10.43.xx.xx:33907 (size: 147.0 B, free: 2.5 GB)
Web UI does not have the use of memory，
||Executor ID||Address||Status||RDD Blocks||Storage Memory||Disk Used||Cores||Active Tasks||Failed Tasks||Complete Tasks||Total Tasks||Task Time (GC Time)||Input||Shuffle Read||Shuffle Write||Logs||Thread Dump||
|0|xxx:38399|Active|0|0.0 B / 2.7 GB|0.0 B|1|0|0|2|2|4 s (0.4 s)|8 B|0.0 B|0.0 B| | |
|driver|xxx:47936|Active|0|0.0 B / 384.1 MB|0.0 B|0|0|0|0|0|0.0 ms (0.0 ms)|0.0 B|0.0 B|0.0 B| | |
|1|xxx:47414|Active|0|0.0 B / 2.7 GB|0.0 B|1|0|0|0|0|0.0 ms (0.0 ms)|0.0 B|0.0 B|0.0 B| | |
|2|xxx:33907|Active|0|0.0 B / 2.7 GB|0.0 B|1|0|0|2|2|4 s (0.2 s)|4 B|0.0 B|0.0 B| | |
was:
The amount of memory used by the broadcast variable is not synchronized to the UI display，
spark-sql> select /*+ broadcast(a)*/ a.id,b.id from a join b on a.id = b.id;
View the app's driver log：
2019-01-25 16:45:23,726 INFO org.apache.spark.storage.BlockManagerInfo: Added broadcast_4_piece0 in memory on 10.43.xx.xx:33907 (size: 6.6 KB, free: 2.5 GB)
2019-01-25 16:45:23,727 INFO org.apache.spark.storage.BlockManagerInfo: Added broadcast_4_piece0 in memory on 10.43.xx.xx:38399 (size: 6.6 KB, free: 2.5 GB)
2019-01-25 16:45:23,745 INFO org.apache.spark.storage.BlockManagerInfo: Added broadcast_3_piece0 in memory on 10.43.xx.xx:33907 (size: 32.1 KB, free: 2.5 GB)
2019-01-25 16:45:23,749 INFO org.apache.spark.storage.BlockManagerInfo: Added broadcast_3_piece0 in memory on 10.43.xx.xx:38399 (size: 32.1 KB, free: 2.5 GB)
2019-01-25 16:45:23,838 INFO org.apache.spark.storage.BlockManagerInfo: Added broadcast_2_piece0 in memory on 10.43.xx.xx:38399 (size: 147.0 B, free: 2.5 GB)
2019-01-25 16:45:23,840 INFO org.apache.spark.storage.BlockManagerInfo: Added broadcast_2_piece0 in memory on 10.43.xx.xx:33907 (size: 147.0 B, free: 2.5 GB)
Web UI does not have the use of memory，
||Executor ID||Address||Status||RDD Blocks||Storage Memory||Disk Used||Cores||Active Tasks||Failed Tasks||Complete Tasks||Total Tasks||Task Time (GC Time)||Input||Shuffle Read||Shuffle Write||Logs||Thread Dump||
|0|xxx:38399|Active|0|0.0 B / 2.7 GB|0.0 B|1|0|0|2|2|4 s (0.4 s)|8 B|0.0 B|0.0 B|[stdout|http://10.43.183.120:18085/logPage/?appId=app-20190125164426-0003&executorId=0&logType=stdout]
[stderr|http://10.43.183.120:18085/logPage/?appId=app-20190125164426-0003&executorId=0&logType=stderr]|[Thread Dump|http://10.43.183.121:4040/executors/threadDump/?executorId=0]|
|driver|xxx:47936|Active|0|0.0 B / 384.1 MB|0.0 B|0|0|0|0|0|0.0 ms (0.0 ms)|0.0 B|0.0 B|0.0 B| |[Thread Dump|http://10.43.183.121:4040/executors/threadDump/?executorId=driver]|
|1|xxx:47414|Active|0|0.0 B / 2.7 GB|0.0 B|1|0|0|0|0|0.0 ms (0.0 ms)|0.0 B|0.0 B|0.0 B|[stdout|http://10.43.183.121:18085/logPage/?appId=app-20190125164426-0003&executorId=1&logType=stdout]
[stderr|http://10.43.183.121:18085/logPage/?appId=app-20190125164426-0003&executorId=1&logType=stderr]|[Thread Dump|http://10.43.183.121:4040/executors/threadDump/?executorId=1]|
|2|xxx:33907|Active|0|0.0 B / 2.7 GB|0.0 B|1|0|0|2|2|4 s (0.2 s)|4 B|0.0 B|0.0 B|[stdout|http://10.43.183.122:18085/logPage/?appId=app-20190125164426-0003&executorId=2&logType=stdout]
[stderr|http://10.43.183.122:18085/logPage/?appId=app-20190125164426-0003&executorId=2&logType=stderr]|[Thread Dump|http://10.43.183.121:4040/executors/threadDump/?executorId=2]|
> Synchronize the amount of memory used by the broadcast variable to the UI display
> -----------------------------------------------------------------------------------
>
> Key: SPARK-26726
> URL: https://issues.apache.org/jira/browse/SPARK-26726
> Project: Spark
> Issue Type: Bug
> Components: Spark Core
> Affects Versions: 2.4.0
> Reporter: hantiantian
> Assignee: hantiantian
> Priority: Major
> Fix For: 2.3.3, 2.4.1, 3.0.0
>
>
> The amount of memory used by the broadcast variable is not synchronized to the UI display，
> spark-sql> select /*+ broadcast(a)*/ a.id,b.id from a join b on a.id = b.id;
> View the app's driver log：
> 2019-01-25 16:45:23,726 INFO org.apache.spark.storage.BlockManagerInfo: Added broadcast_4_piece0 in memory on 10.43.xx.xx:33907 (size: 6.6 KB, free: 2.5 GB)
> 2019-01-25 16:45:23,727 INFO org.apache.spark.storage.BlockManagerInfo: Added broadcast_4_piece0 in memory on 10.43.xx.xx:38399 (size: 6.6 KB, free: 2.5 GB)
> 2019-01-25 16:45:23,745 INFO org.apache.spark.storage.BlockManagerInfo: Added broadcast_3_piece0 in memory on 10.43.xx.xx:33907 (size: 32.1 KB, free: 2.5 GB)
> 2019-01-25 16:45:23,749 INFO org.apache.spark.storage.BlockManagerInfo: Added broadcast_3_piece0 in memory on 10.43.xx.xx:38399 (size: 32.1 KB, free: 2.5 GB)
> 2019-01-25 16:45:23,838 INFO org.apache.spark.storage.BlockManagerInfo: Added broadcast_2_piece0 in memory on 10.43.xx.xx:38399 (size: 147.0 B, free: 2.5 GB)
> 2019-01-25 16:45:23,840 INFO org.apache.spark.storage.BlockManagerInfo: Added broadcast_2_piece0 in memory on 10.43.xx.xx:33907 (size: 147.0 B, free: 2.5 GB)
>
> Web UI does not have the use of memory，
> ||Executor ID||Address||Status||RDD Blocks||Storage Memory||Disk Used||Cores||Active Tasks||Failed Tasks||Complete Tasks||Total Tasks||Task Time (GC Time)||Input||Shuffle Read||Shuffle Write||Logs||Thread Dump||
> |0|xxx:38399|Active|0|0.0 B / 2.7 GB|0.0 B|1|0|0|2|2|4 s (0.4 s)|8 B|0.0 B|0.0 B| | |
> |driver|xxx:47936|Active|0|0.0 B / 384.1 MB|0.0 B|0|0|0|0|0|0.0 ms (0.0 ms)|0.0 B|0.0 B|0.0 B| | |
> |1|xxx:47414|Active|0|0.0 B / 2.7 GB|0.0 B|1|0|0|0|0|0.0 ms (0.0 ms)|0.0 B|0.0 B|0.0 B| | |
> |2|xxx:33907|Active|0|0.0 B / 2.7 GB|0.0 B|1|0|0|2|2|4 s (0.2 s)|4 B|0.0 B|0.0 B| | |
>
>
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@spark.apache.org
For additional commands, e-mail: issues-help@spark.apache.org