[CASSANDRA-10536] Batch statements with multiple updates to partition error when table is indexed

• Type: Bug
• Status: Resolved
• Priority: Normal
• Resolution: Fixed
• Fix Version/s: 3.0.0 rc2
• Component/s: Legacy/CQL
• Severity: Normal
• Assignee: Sam Tunnicliffe
• Reporter: Tom Hobbs

Authors

Sam Tunnicliffe

Reviewers

Aleksey Yeschenko

• Created: 15/Oct/15 21:04
• Updated: 16/Apr/19 09:30
• Resolved: 16/Oct/15 17:05

Description

If a BATCH statement contains multiple UPDATE statements that update the same partition, and a secondary index exists on that table, the batch statement will error:

ServerError: <ErrorMessage code=0000 [Server error] message="java.lang.IllegalStateException: An update should not be written again once it has been read">

with the following traceback in the logs:

ERROR 20:53:46 Unexpected exception during request
java.lang.IllegalStateException: An update should not be written again once it has been read
	at org.apache.cassandra.db.partitions.PartitionUpdate.assertNotBuilt(PartitionUpdate.java:504) ~[main/:na]
	at org.apache.cassandra.db.partitions.PartitionUpdate.add(PartitionUpdate.java:535) ~[main/:na]
	at org.apache.cassandra.cql3.statements.UpdateStatement.addUpdateForKey(UpdateStatement.java:96) ~[main/:na]
	at org.apache.cassandra.cql3.statements.ModificationStatement.addUpdates(ModificationStatement.java:667) ~[main/:na]
	at org.apache.cassandra.cql3.statements.BatchStatement.getMutations(BatchStatement.java:234) ~[main/:na]
	at org.apache.cassandra.cql3.statements.BatchStatement.execute(BatchStatement.java:335) ~[main/:na]
	at org.apache.cassandra.cql3.statements.BatchStatement.execute(BatchStatement.java:321) ~[main/:na]
	at org.apache.cassandra.cql3.statements.BatchStatement.execute(BatchStatement.java:316) ~[main/:na]
	at org.apache.cassandra.cql3.QueryProcessor.processStatement(QueryProcessor.java:205) ~[main/:na]
	at org.apache.cassandra.cql3.QueryProcessor.processPrepared(QueryProcessor.java:471) ~[main/:na]
	at org.apache.cassandra.cql3.QueryProcessor.processPrepared(QueryProcessor.java:448) ~[main/:na]
	at org.apache.cassandra.transport.messages.ExecuteMessage.execute(ExecuteMessage.java:130) ~[main/:na]
	at org.apache.cassandra.transport.Message$Dispatcher.channelRead0(Message.java:507) [main/:na]
	at org.apache.cassandra.transport.Message$Dispatcher.channelRead0(Message.java:401) [main/:na]
	at io.netty.channel.SimpleChannelInboundHandler.channelRead(SimpleChannelInboundHandler.java:105) [netty-all-4.0.23.Final.jar:4.0.23.Final]
	at io.netty.channel.AbstractChannelHandlerContext.invokeChannelRead(AbstractChannelHandlerContext.java:333) [netty-all-4.0.23.Final.jar:4.0.23.Final]
	at io.netty.channel.AbstractChannelHandlerContext.access$700(AbstractChannelHandlerContext.java:32) [netty-all-4.0.23.Final.jar:4.0.23.Final]
	at io.netty.channel.AbstractChannelHandlerContext$8.run(AbstractChannelHandlerContext.java:324) [netty-all-4.0.23.Final.jar:4.0.23.Final]
	at java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:511) [na:1.8.0_45]
	at org.apache.cassandra.concurrent.AbstractTracingAwareExecutorService$FutureTask.run(AbstractTracingAwareExecutorService.java:164) [main/:na]
	at org.apache.cassandra.concurrent.SEPWorker.run(SEPWorker.java:105) [main/:na]
	at java.lang.Thread.run(Thread.java:745) [na:1.8.0_45]

This is due to SecondaryIndexManager.validate() triggering a build of the PartitionUpdate (stacktrace from debugging the build() call):

at org.apache.cassandra.db.partitions.PartitionUpdate.build(PartitionUpdate.java:571) [main/:na]
	at org.apache.cassandra.db.partitions.PartitionUpdate.maybeBuild(PartitionUpdate.java:561) [main/:na]
	at org.apache.cassandra.db.partitions.PartitionUpdate.iterator(PartitionUpdate.java:418) [main/:na]
	at org.apache.cassandra.index.internal.CassandraIndex.validateRows(CassandraIndex.java:560) [main/:na]
	at org.apache.cassandra.index.internal.CassandraIndex.validate(CassandraIndex.java:314) [main/:na]
	at org.apache.cassandra.index.SecondaryIndexManager.lambda$validate$75(SecondaryIndexManager.java:642) [main/:na]
	at org.apache.cassandra.index.SecondaryIndexManager$$Lambda$166/1388080038.accept(Unknown Source) [main/:na]
	at java.util.stream.ForEachOps$ForEachOp$OfRef.accept(ForEachOps.java:184) [na:1.8.0_45]
	at java.util.stream.ReferencePipeline$2$1.accept(ReferencePipeline.java:175) [na:1.8.0_45]
	at java.util.concurrent.ConcurrentHashMap$ValueSpliterator.forEachRemaining(ConcurrentHashMap.java:3566) [na:1.8.0_45]
	at java.util.stream.AbstractPipeline.copyInto(AbstractPipeline.java:512) [na:1.8.0_45]
	at java.util.stream.AbstractPipeline.wrapAndCopyInto(AbstractPipeline.java:502) [na:1.8.0_45]
	at java.util.stream.ForEachOps$ForEachOp.evaluateSequential(ForEachOps.java:151) [na:1.8.0_45]
	at java.util.stream.ForEachOps$ForEachOp$OfRef.evaluateSequential(ForEachOps.java:174) [na:1.8.0_45]
	at java.util.stream.AbstractPipeline.evaluate(AbstractPipeline.java:234) [na:1.8.0_45]
	at java.util.stream.ReferencePipeline.forEach(ReferencePipeline.java:418) [na:1.8.0_45]
	at org.apache.cassandra.index.SecondaryIndexManager.validate(SecondaryIndexManager.java:642) [main/:na]
	at org.apache.cassandra.cql3.UpdateParameters.validateIndexedColumns(UpdateParameters.java:105) [main/:na]
	at org.apache.cassandra.cql3.statements.UpdateStatement.addUpdateForKey(UpdateStatement.java:107) [main/:na]
	at org.apache.cassandra.cql3.statements.ModificationStatement.addUpdates(ModificationStatement.java:667) [main/:na]
	at org.apache.cassandra.cql3.statements.BatchStatement.getMutations(BatchStatement.java:234) [main/:na]
	at org.apache.cassandra.cql3.statements.BatchStatement.execute(BatchStatement.java:335) [main/:na]
	at org.apache.cassandra.cql3.statements.BatchStatement.execute(BatchStatement.java:321) [main/:na]
	at org.apache.cassandra.cql3.statements.BatchStatement.execute(BatchStatement.java:316) [main/:na]
	at org.apache.cassandra.cql3.QueryProcessor.processStatement(QueryProcessor.java:205) [main/:na]
	at org.apache.cassandra.cql3.QueryProcessor.processPrepared(QueryProcessor.java:471) [main/:na]
	at org.apache.cassandra.cql3.QueryProcessor.processPrepared(QueryProcessor.java:448) [main/:na]
	at org.apache.cassandra.transport.messages.ExecuteMessage.execute(ExecuteMessage.java:130) [main/:na]
	at org.apache.cassandra.transport.Message$Dispatcher.channelRead0(Message.java:507) [main/:na]
	at org.apache.cassandra.transport.Message$Dispatcher.channelRead0(Message.java:401) [main/:na]
	at io.netty.channel.SimpleChannelInboundHandler.channelRead(SimpleChannelInboundHandler.java:105) [netty-all-4.0.23.Final.jar:4.0.23.Final]
	at io.netty.channel.AbstractChannelHandlerContext.invokeChannelRead(AbstractChannelHandlerContext.java:333) [netty-all-4.0.23.Final.jar:4.0.23.Final]
	at io.netty.channel.AbstractChannelHandlerContext.access$700(AbstractChannelHandlerContext.java:32) [netty-all-4.0.23.Final.jar:4.0.23.Final]
	at io.netty.channel.AbstractChannelHandlerContext$8.run(AbstractChannelHandlerContext.java:324) [netty-all-4.0.23.Final.jar:4.0.23.Final]
	at java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:511) [na:1.8.0_45]
	at org.apache.cassandra.concurrent.AbstractTracingAwareExecutorService$FutureTask.run(AbstractTracingAwareExecutorService.java:164) [main/:na]
	at org.apache.cassandra.concurrent.SEPWorker.run(SEPWorker.java:105) [main/:na]
	at java.lang.Thread.run(Thread.java:745) [na:1.8.0_45]

You can reproduce the issue with a table and index like this:

CREATE TABLE foo (a int, b int, c int, d int, PRIMARY KEY (a, b));
CREATE INDEX ON foo(c);

and a batch update like this:

BEGIN BATCH
UPDATE foo SET c = 0 WHERE a = 0 AND b = 0;
UPDATE foo SET d = 0 WHERE a = 0 AND b = 0;
APPLY BATCH

Assigning slebresne for now, since he probably has the best idea of how to avoid this.