[BEAM-981] Not possible to directly submit a pipeline on spark cluster

• Type: Bug
• Status: Resolved
• Priority: P2
• Resolution: Cannot Reproduce
• Affects Version/s: 0.6.0
• Fix Version/s: Not applicable
• Component/s: runner-spark
• Assignee: Unassigned
• Reporter: Jean-Baptiste Onofré
• Created: 15/Nov/16 16:52
• Updated: 16/May/20 14:18
• Resolved: 05/Mar/20 12:05

Description

It's not possible to directly run a pipeline on the spark runner (for instance using mvn exec:java. It fails with:

[appclient-register-master-threadpool-0] INFO org.apache.spark.deploy.client.AppClient$ClientEndpoint - Connecting to master spark://10.200.118.197:7077...
[shuffle-client-0] ERROR org.apache.spark.network.client.TransportClient - Failed to send RPC 6813731522650020739 to /10.200.118.197:7077: java.lang.AbstractMethodError: org.apache.spark.network.protocol.MessageWithHeader.touch(Ljava/lang/Object;)Lio/netty/util/ReferenceCounted;
java.lang.AbstractMethodError: org.apache.spark.network.protocol.MessageWithHeader.touch(Ljava/lang/Object;)Lio/netty/util/ReferenceCounted;
        at io.netty.util.ReferenceCountUtil.touch(ReferenceCountUtil.java:73)
        at io.netty.channel.DefaultChannelPipeline.touch(DefaultChannelPipeline.java:107)
        at io.netty.channel.AbstractChannelHandlerContext.write(AbstractChannelHandlerContext.java:820)
        at io.netty.channel.AbstractChannelHandlerContext.write(AbstractChannelHandlerContext.java:733)
        at io.netty.handler.codec.MessageToMessageEncoder.write(MessageToMessageEncoder.java:111)
        at io.netty.channel.AbstractChannelHandlerContext.invokeWrite0(AbstractChannelHandlerContext.java:748)
        at io.netty.channel.AbstractChannelHandlerContext.invokeWrite(AbstractChannelHandlerContext.java:740)
        at io.netty.channel.AbstractChannelHandlerContext.write(AbstractChannelHandlerContext.java:826)
        at io.netty.channel.AbstractChannelHandlerContext.write(AbstractChannelHandlerContext.java:733)
        at io.netty.handler.timeout.IdleStateHandler.write(IdleStateHandler.java:284)
        at io.netty.channel.AbstractChannelHandlerContext.invokeWrite0(AbstractChannelHandlerContext.java:748)
        at io.netty.channel.AbstractChannelHandlerContext.invokeWrite(AbstractChannelHandlerContext.java:740)
        at io.netty.channel.AbstractChannelHandlerContext.access$1900(AbstractChannelHandlerContext.java:38)
        at io.netty.channel.AbstractChannelHandlerContext$AbstractWriteTask.write(AbstractChannelHandlerContext.java:1101)
        at io.netty.channel.AbstractChannelHandlerContext$WriteAndFlushTask.write(AbstractChannelHandlerContext.java:1148)
        at io.netty.channel.AbstractChannelHandlerContext$AbstractWriteTask.run(AbstractChannelHandlerContext.java:1090)
        at io.netty.util.concurrent.SingleThreadEventExecutor.safeExecute(SingleThreadEventExecutor.java:451)
        at io.netty.util.concurrent.SingleThreadEventExecutor.runAllTasks(SingleThreadEventExecutor.java:418)
        at io.netty.channel.nio.NioEventLoop.run(NioEventLoop.java:401)
        at io.netty.util.concurrent.SingleThreadEventExecutor$5.run(SingleThreadEventExecutor.java:877)
        at java.lang.Thread.run(Thread.java:745)
[appclient-register-master-threadpool-0] WARN org.apache.spark.deploy.client.AppClient$ClientEndpoint - Failed to connect to master 10.200.118.197:7077
java.io.IOException: Failed to send RPC 6813731522650020739 to /10.200.118.197:7077: java.lang.AbstractMethodError: org.apache.spark.network.protocol.MessageWithHeader.touch(Ljava/lang/Object;)Lio/netty/util/ReferenceCounted;
        at org.apache.spark.network.client.TransportClient$3.operationComplete(TransportClient.java:239)
        at org.apache.spark.network.client.TransportClient$3.operationComplete(TransportClient.java:226)
        at io.netty.util.concurrent.DefaultPromise.notifyListener0(DefaultPromise.java:514)
        at io.netty.util.concurrent.DefaultPromise.notifyListeners0(DefaultPromise.java:507)
        at io.netty.util.concurrent.DefaultPromise.notifyListenersNow(DefaultPromise.java:486)
        at io.netty.util.concurrent.DefaultPromise.notifyListeners(DefaultPromise.java:427)
        at io.netty.util.concurrent.DefaultPromise.tryFailure(DefaultPromise.java:129)
        at io.netty.channel.AbstractChannelHandlerContext.notifyOutboundHandlerException(AbstractChannelHandlerContext.java:845)
        at io.netty.channel.AbstractChannelHandlerContext.invokeWrite0(AbstractChannelHandlerContext.java:750)
        at io.netty.channel.AbstractChannelHandlerContext.invokeWrite(AbstractChannelHandlerContext.java:740)
        at io.netty.channel.AbstractChannelHandlerContext.write(AbstractChannelHandlerContext.java:826)
        at io.netty.channel.AbstractChannelHandlerContext.write(AbstractChannelHandlerContext.java:733)
        at io.netty.handler.timeout.IdleStateHandler.write(IdleStateHandler.java:284)
        at io.netty.channel.AbstractChannelHandlerContext.invokeWrite0(AbstractChannelHandlerContext.java:748)
        at io.netty.channel.AbstractChannelHandlerContext.invokeWrite(AbstractChannelHandlerContext.java:740)
        at io.netty.channel.AbstractChannelHandlerContext.access$1900(AbstractChannelHandlerContext.java:38)
        at io.netty.channel.AbstractChannelHandlerContext$AbstractWriteTask.write(AbstractChannelHandlerContext.java:1101)
        at io.netty.channel.AbstractChannelHandlerContext$WriteAndFlushTask.write(AbstractChannelHandlerContext.java:1148)
        at io.netty.channel.AbstractChannelHandlerContext$AbstractWriteTask.run(AbstractChannelHandlerContext.java:1090)
        at io.netty.util.concurrent.SingleThreadEventExecutor.safeExecute(SingleThreadEventExecutor.java:451)
        at io.netty.util.concurrent.SingleThreadEventExecutor.runAllTasks(SingleThreadEventExecutor.java:418)
        at io.netty.channel.nio.NioEventLoop.run(NioEventLoop.java:401)
        at io.netty.util.concurrent.SingleThreadEventExecutor$5.run(SingleThreadEventExecutor.java:877)
        at java.lang.Thread.run(Thread.java:745)
Caused by: java.lang.AbstractMethodError: org.apache.spark.network.protocol.MessageWithHeader.touch(Ljava/lang/Object;)Lio/netty/util/ReferenceCounted;
        at io.netty.util.ReferenceCountUtil.touch(ReferenceCountUtil.java:73)
        at io.netty.channel.DefaultChannelPipeline.touch(DefaultChannelPipeline.java:107)
        at io.netty.channel.AbstractChannelHandlerContext.write(AbstractChannelHandlerContext.java:820)
        at io.netty.channel.AbstractChannelHandlerContext.write(AbstractChannelHandlerContext.java:733)
        at io.netty.handler.codec.MessageToMessageEncoder.write(MessageToMessageEncoder.java:111)
        at io.netty.channel.AbstractChannelHandlerContext.invokeWrite0(AbstractChannelHandlerContext.java:748)
        ... 15 more

It looks like a conflict between the Netty version used in Spark and the one in Beam (just guessing).

The workaround is to use spark-submit.

Issue Links

• is related to: BEAM-1603 Enable programmatic execution of spark pipelines.
• is required by: BEAM-3371 Add ability to stage directories with compiled classes to Spark
• links to: GitHub Pull Request #2997
• links to: GitHub Pull Request #4246

Sub-Tasks

• Support adding jars to spark runner: Open
• Document use of adding jars to spark runner: Open