[
https://issues.apache.org/jira/browse/DRILL-5032?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15673402#comment-15673402
]
Serhii Harnyk commented on DRILL-5032:
--------------------------------------
List of columns in every hive partition increases size of serialized physical
plan that causes OOM.
As Hive allows every partition to have its own set of columns now system
checks, and if all column sets for all partitions are equal, list of columns
stores only on hive table level.
> Drill query on hive parquet table failed with OutOfMemoryError: Java heap
> space
> -------------------------------------------------------------------------------
>
> Key: DRILL-5032
> URL: https://issues.apache.org/jira/browse/DRILL-5032
> Project: Apache Drill
> Issue Type: Bug
> Components: Functions - Hive
> Affects Versions: 1.8.0
> Reporter: Serhii Harnyk
> Assignee: Serhii Harnyk
>
> Following query on hive parquet table failed with OOM Java heap space:
> {code}
> select distinct(businessdate) from vmdr_trades where trade_date='2016-04-12'
> 2016-08-31 08:02:03,597 [283938c3-fde8-0fc6-37e1-9a568c7f5913:foreman] INFO
> o.a.drill.exec.work.foreman.Foreman - Query text for query id
> 283938c3-fde8-0fc6-37e1-9a568c7f5913: select distinct(businessdate) from
> vmdr_trades where trade_date='2016-04-12'
> 2016-08-31 08:05:58,502 [283938c3-fde8-0fc6-37e1-9a568c7f5913:foreman] INFO
> o.a.d.e.p.l.partition.PruneScanRule - Beginning partition pruning, pruning
> class:
> org.apache.drill.exec.planner.sql.logical.HivePushPartitionFilterIntoScan$2
> 2016-08-31 08:05:58,506 [283938c3-fde8-0fc6-37e1-9a568c7f5913:foreman] INFO
> o.a.d.e.p.l.partition.PruneScanRule - Total elapsed time to build and analyze
> filter tree: 1 ms
> 2016-08-31 08:05:58,506 [283938c3-fde8-0fc6-37e1-9a568c7f5913:foreman] INFO
> o.a.d.e.p.l.partition.PruneScanRule - No conditions were found eligible for
> partition pruning.Total pruning elapsed time: 3 ms
> 2016-08-31 08:05:58,663 [283938c3-fde8-0fc6-37e1-9a568c7f5913:foreman] INFO
> o.a.d.e.p.l.partition.PruneScanRule - Beginning partition pruning, pruning
> class:
> org.apache.drill.exec.planner.sql.logical.HivePushPartitionFilterIntoScan$2
> 2016-08-31 08:05:58,663 [283938c3-fde8-0fc6-37e1-9a568c7f5913:foreman] INFO
> o.a.d.e.p.l.partition.PruneScanRule - Total elapsed time to build and analyze
> filter tree: 0 ms
> 2016-08-31 08:05:58,663 [283938c3-fde8-0fc6-37e1-9a568c7f5913:foreman] INFO
> o.a.d.e.p.l.partition.PruneScanRule - No conditions were found eligible for
> partition pruning.Total pruning elapsed time: 0 ms
> 2016-08-31 08:05:58,664 [283938c3-fde8-0fc6-37e1-9a568c7f5913:foreman] INFO
> o.a.d.e.p.l.partition.PruneScanRule - Beginning partition pruning, pruning
> class:
> org.apache.drill.exec.planner.sql.logical.HivePushPartitionFilterIntoScan$1
> 2016-08-31 08:05:58,665 [283938c3-fde8-0fc6-37e1-9a568c7f5913:foreman] INFO
> o.a.d.e.p.l.partition.PruneScanRule - Total elapsed time to build and analyze
> filter tree: 0 ms
> 2016-08-31 08:05:58,665 [283938c3-fde8-0fc6-37e1-9a568c7f5913:foreman] INFO
> o.a.d.e.p.l.partition.PruneScanRule - No conditions were found eligible for
> partition pruning.Total pruning elapsed time: 0 ms
> 2016-08-31 08:09:42,355 [283938c3-fde8-0fc6-37e1-9a568c7f5913:foreman] ERROR
> o.a.drill.common.CatastrophicFailure - Catastrophic Failure Occurred,
> exiting. Information message: Unable to handle out of memory condition in
> Foreman.
> java.lang.OutOfMemoryError: Java heap space
> at java.util.Arrays.copyOf(Arrays.java:3332) ~[na:1.8.0_74]
> at
> java.lang.AbstractStringBuilder.expandCapacity(AbstractStringBuilder.java:137)
> ~[na:1.8.0_74]
> at
> java.lang.AbstractStringBuilder.ensureCapacityInternal(AbstractStringBuilder.java:121)
> ~[na:1.8.0_74]
> at
> java.lang.AbstractStringBuilder.append(AbstractStringBuilder.java:421)
> ~[na:1.8.0_74]
> at java.lang.StringBuilder.append(StringBuilder.java:136)
> ~[na:1.8.0_74]
> at java.lang.StringBuilder.append(StringBuilder.java:76)
> ~[na:1.8.0_74]
> at
> java.lang.AbstractStringBuilder.append(AbstractStringBuilder.java:457)
> ~[na:1.8.0_74]
> at java.lang.StringBuilder.append(StringBuilder.java:166)
> ~[na:1.8.0_74]
> at java.lang.StringBuilder.append(StringBuilder.java:76)
> ~[na:1.8.0_74]
> at
> com.google.protobuf.TextFormat$TextGenerator.write(TextFormat.java:538)
> ~[protobuf-java-2.5.0.jar:na]
> at
> com.google.protobuf.TextFormat$TextGenerator.print(TextFormat.java:526)
> ~[protobuf-java-2.5.0.jar:na]
> at
> com.google.protobuf.TextFormat$Printer.printFieldValue(TextFormat.java:389)
> ~[protobuf-java-2.5.0.jar:na]
> at
> com.google.protobuf.TextFormat$Printer.printSingleField(TextFormat.java:327)
> ~[protobuf-java-2.5.0.jar:na]
> at
> com.google.protobuf.TextFormat$Printer.printField(TextFormat.java:286)
> ~[protobuf-java-2.5.0.jar:na]
> at com.google.protobuf.TextFormat$Printer.print(TextFormat.java:273)
> ~[protobuf-java-2.5.0.jar:na]
> at
> com.google.protobuf.TextFormat$Printer.access$400(TextFormat.java:248)
> ~[protobuf-java-2.5.0.jar:na]
> at com.google.protobuf.TextFormat.print(TextFormat.java:71)
> ~[protobuf-java-2.5.0.jar:na]
> at com.google.protobuf.TextFormat.printToString(TextFormat.java:118)
> ~[protobuf-java-2.5.0.jar:na]
> at
> com.google.protobuf.AbstractMessage.toString(AbstractMessage.java:106)
> ~[protobuf-java-2.5.0.jar:na]
> at
> org.apache.drill.exec.planner.fragment.SimpleParallelizer.generateWorkUnit(SimpleParallelizer.java:395)
> ~[drill-java-exec-1.6.0.jar:1.6.0]
> at
> org.apache.drill.exec.planner.fragment.SimpleParallelizer.getFragments(SimpleParallelizer.java:134)
> ~[drill-java-exec-1.6.0.jar:1.6.0]
> at
> org.apache.drill.exec.work.foreman.Foreman.getQueryWorkUnit(Foreman.java:516)
> ~[drill-java-exec-1.6.0.jar:1.6.0]
> at
> org.apache.drill.exec.work.foreman.Foreman.runPhysicalPlan(Foreman.java:403)
> ~[drill-java-exec-1.6.0.jar:1.6.0]
> at
> org.apache.drill.exec.work.foreman.Foreman.runSQL(Foreman.java:929)
> ~[drill-java-exec-1.6.0.jar:1.6.0]
> at org.apache.drill.exec.work.foreman.Foreman.run(Foreman.java:251)
> ~[drill-java-exec-1.6.0.jar:1.6.0]
> at
> java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1142)
> [na:1.8.0_74]
> at
> java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:617)
> [na:1.8.0_74]
> at java.lang.Thread.run(Thread.java:745) [na:1.8.0_74]
> 2016-08-31 08:09:43,358 [Drillbit-ShutdownHook#0] INFO
> o.apache.drill.exec.server.Drillbit - Received shutdown request.
> 2016-08-31 08:09:49,385 [BitServer-3] INFO
> o.a.d.exec.rpc.control.ControlClient - Channel closed /162.111.92.29:33973
> <--> /162.111.92.29:31011.
> 2016-08-31 08:09:50,388 [pool-6-thread-2] INFO
> o.a.drill.exec.rpc.data.DataServer - closed eventLoopGroup
> io.netty.channel.epoll.EpollEventLoopGroup@268b55a9 in 1007 ms
> 2016-08-31 08:09:50,389 [pool-6-thread-2] INFO
> o.a.drill.exec.service.ServiceEngine - closed dataPool in 1008 ms
> {code}
> The Drill cluster having 16 nodes and current drill direct memory is 16GB
> and heap memory 20GB. table vmdr_trades is ~19GB and having 750+ partition
> and each partition having one parquet file roughly.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
