[
https://issues.apache.org/jira/browse/SPARK-33620?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Vladislav Sterkhov updated SPARK-33620:
---------------------------------------
Description:
Hello i have problem with big memory used ~2000gb hdfs stack. With 300gb memory
used task starting and complete, but we need use unlimited stack. Please help
!VlwWJ.png|width=644,height=150!
!mgg1s.png|width=651,height=182!
This my code:
{{var filteredRDD = sparkContext.emptyRDD[String]
for (path<- pathBuffer){
val someRDD = sparkContext.textFile(path)
if (isValidRDD(someRDD))
filteredRDD = filteredRDD.++(someRDD.filter(row =>\{...})
}
hiveService.insertRDD(filteredRDD.repartition(10), outTable)}}
been other way. When i got StackOverflowError after many iteration spark
{{java.lang.StackOverflowError
at java.io.ObjectInputStream$PeekInputStream.peek(ObjectInputStream.java:2303)
at
java.io.ObjectInputStream$BlockDataInputStream.peek(ObjectInputStream.java:2596)
at
java.io.ObjectInputStream$BlockDataInputStream.peekByte(ObjectInputStream.java:2606)
at java.io.ObjectInputStream.readObject0(ObjectInputStream.java:1319)
at java.io.ObjectInputStream.defaultReadFields(ObjectInputStream.java:2000)
at java.io.ObjectInputStream.readSerialData(ObjectInputStream.java:1924)
at java.io.ObjectInputStream.readOrdinaryObject(ObjectInputStream.java:1801)
at java.io.ObjectInputStream.readObject0(ObjectInputStream.java:1351)
at java.io.ObjectInputStream.defaultReadFields(ObjectInputStream.java:2000)
at java.io.ObjectInputStream.readSerialData(ObjectInputStream.java:1924)
at java.io.ObjectInputStream.readOrdinaryObject(ObjectInputStream.java:1801)
at java.io.ObjectInputStream.readObject0(ObjectInputStream.java:1351)
at java.io.ObjectInputStream.readArray(ObjectInputStream.java:1707)
at java.io.ObjectInputStream.readObject0(ObjectInputStream.java:1345)
at java.io.ObjectInputStream.defaultReadFields(ObjectInputStream.java:2000)
at java.io.ObjectInputStream.readSerialData(ObjectInputStream.java:1924)
at java.io.ObjectInputStream.readOrdinaryObject(ObjectInputStream.java:1801)}}
\{{}}
\{{}}
How i must build my code with repartitional and persist\coalesce for to nodes
not crashes?
I tried to rebuild the program in different ways, transferring repartitioning
and saving in memory / disk inside the loop, installed a large number of
partitions - 200.
The program either hangs on the “repartition” stage or crashes into error code
143 (outOfMemory), throwing a stackOverflowError in a strange way.
was:
Hello i have problem with big memory used ~2000gb hdfs stack. With 300gb memory
used task starting and complete, but we need use unlimited stack. Please help
!VlwWJ.png|width=644,height=150!
!mgg1s.png|width=651,height=182!
This my code:
{{var filteredRDD = sparkContext.emptyRDD[String]
for (path<- pathBuffer)\{
val someRDD = sparkContext.textFile(path)
if (isValidRDD(someRDD))
filteredRDD = filteredRDD.++(someRDD.filter(row => {...})
}
hiveService.insertRDD(filteredRDD.repartition(10), outTable)}}
{{}}
{{}}
been other way. When i got StackOverflowError after many iteration spark
{{java.lang.StackOverflowError
at
java.io.ObjectInputStream$PeekInputStream.peek(ObjectInputStream.java:2303)
at
java.io.ObjectInputStream$BlockDataInputStream.peek(ObjectInputStream.java:2596)
at
java.io.ObjectInputStream$BlockDataInputStream.peekByte(ObjectInputStream.java:2606)
at java.io.ObjectInputStream.readObject0(ObjectInputStream.java:1319)
at java.io.ObjectInputStream.defaultReadFields(ObjectInputStream.java:2000)
at java.io.ObjectInputStream.readSerialData(ObjectInputStream.java:1924)
at java.io.ObjectInputStream.readOrdinaryObject(ObjectInputStream.java:1801)
at java.io.ObjectInputStream.readObject0(ObjectInputStream.java:1351)
at java.io.ObjectInputStream.defaultReadFields(ObjectInputStream.java:2000)
at java.io.ObjectInputStream.readSerialData(ObjectInputStream.java:1924)
at java.io.ObjectInputStream.readOrdinaryObject(ObjectInputStream.java:1801)
at java.io.ObjectInputStream.readObject0(ObjectInputStream.java:1351)
at java.io.ObjectInputStream.readArray(ObjectInputStream.java:1707)
at java.io.ObjectInputStream.readObject0(ObjectInputStream.java:1345)
at java.io.ObjectInputStream.defaultReadFields(ObjectInputStream.java:2000)
at java.io.ObjectInputStream.readSerialData(ObjectInputStream.java:1924)
at
java.io.ObjectInputStream.readOrdinaryObject(ObjectInputStream.java:1801)}}
{{}}
{{}}
How i must build my code with repartitional and persist\coalesce for to nodes
not crashes?
I tried to rebuild the program in different ways, transferring repartitioning
and saving in memory / disk inside the loop, installed a large number of
partitions - 200.
The program either hangs on the “repartition” stage or crashes into error code
143 (outOfMemory), throwing a stackOverflowError in a strange way.
> Task not started after filtering
> --------------------------------
>
> Key: SPARK-33620
> URL: https://issues.apache.org/jira/browse/SPARK-33620
> Project: Spark
> Issue Type: Question
> Components: Spark Core
> Affects Versions: 2.4.7
> Reporter: Vladislav Sterkhov
> Priority: Major
> Attachments: VlwWJ.png, mgg1s.png
>
>
> Hello i have problem with big memory used ~2000gb hdfs stack. With 300gb
> memory used task starting and complete, but we need use unlimited stack.
> Please help
>
> !VlwWJ.png|width=644,height=150!
>
> !mgg1s.png|width=651,height=182!
>
> This my code:
> {{var filteredRDD = sparkContext.emptyRDD[String]
> for (path<- pathBuffer){
> val someRDD = sparkContext.textFile(path)
> if (isValidRDD(someRDD))
> filteredRDD = filteredRDD.++(someRDD.filter(row =>\{...})
> }
> hiveService.insertRDD(filteredRDD.repartition(10), outTable)}}
>
> been other way. When i got StackOverflowError after many iteration spark
>
> {{java.lang.StackOverflowError
> at
> java.io.ObjectInputStream$PeekInputStream.peek(ObjectInputStream.java:2303)
> at
> java.io.ObjectInputStream$BlockDataInputStream.peek(ObjectInputStream.java:2596)
> at
> java.io.ObjectInputStream$BlockDataInputStream.peekByte(ObjectInputStream.java:2606)
> at java.io.ObjectInputStream.readObject0(ObjectInputStream.java:1319)
> at java.io.ObjectInputStream.defaultReadFields(ObjectInputStream.java:2000)
> at java.io.ObjectInputStream.readSerialData(ObjectInputStream.java:1924)
> at java.io.ObjectInputStream.readOrdinaryObject(ObjectInputStream.java:1801)
> at java.io.ObjectInputStream.readObject0(ObjectInputStream.java:1351)
> at java.io.ObjectInputStream.defaultReadFields(ObjectInputStream.java:2000)
> at java.io.ObjectInputStream.readSerialData(ObjectInputStream.java:1924)
> at java.io.ObjectInputStream.readOrdinaryObject(ObjectInputStream.java:1801)
> at java.io.ObjectInputStream.readObject0(ObjectInputStream.java:1351)
> at java.io.ObjectInputStream.readArray(ObjectInputStream.java:1707)
> at java.io.ObjectInputStream.readObject0(ObjectInputStream.java:1345)
> at java.io.ObjectInputStream.defaultReadFields(ObjectInputStream.java:2000)
> at java.io.ObjectInputStream.readSerialData(ObjectInputStream.java:1924)
> at
> java.io.ObjectInputStream.readOrdinaryObject(ObjectInputStream.java:1801)}}
> \{{}}
> \{{}}
> How i must build my code with repartitional and persist\coalesce for to
> nodes not crashes?
> I tried to rebuild the program in different ways, transferring repartitioning
> and saving in memory / disk inside the loop, installed a large number of
> partitions - 200.
> The program either hangs on the “repartition” stage or crashes into error
> code 143 (outOfMemory), throwing a stackOverflowError in a strange way.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org
For additional commands, e-mail: issues-h...@spark.apache.org
