[FLINK-31675] Deadlock in AWS Connectors following content-length AWS SDK exception

• Type: Bug
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 1.17.0, 1.16.1, 1.15.4
• Fix Version/s: aws-connector-3.1.0, aws-connector-4.2.0
• Component/s: Connectors / AWS
• Labels: pull-request-available
• Assignee: Antonio Vespoli
• Reporter: Antonio Vespoli
• Created: 30/Mar/23 14:34
• Updated: 26/Apr/23 14:33
• Resolved: 26/Apr/23 14:33

Description

Connector calls to AWS services can hang on a canceled future following a content-length mismatch that isn't handled gracefully by the SDK:

 

org.apache.flink.kinesis.shaded.software.amazon.awssdk.http.nio.netty.internal.FutureCancelledException: java.io.IOException: Response had content-length of 31 bytes, but only received 0 bytes before the connection was closed.
at org.apache.flink.kinesis.shaded.software.amazon.awssdk.http.nio.netty.internal.NettyRequestExecutor.lambda$null$3(NettyRequestExecutor.java:136)
at org.apache.flink.kinesis.shaded.io.netty.util.concurrent.PromiseTask.runTask(PromiseTask.java:98)
at org.apache.flink.kinesis.shaded.io.netty.util.concurrent.PromiseTask.run(PromiseTask.java:106)
at org.apache.flink.kinesis.shaded.io.netty.util.concurrent.AbstractEventExecutor.safeExecute(AbstractEventExecutor.java:164)
at org.apache.flink.kinesis.shaded.io.netty.util.concurrent.SingleThreadEventExecutor.runAllTasks(SingleThreadEventExecutor.java:469)
at org.apache.flink.kinesis.shaded.io.netty.channel.nio.NioEventLoop.run(NioEventLoop.java:500)
at org.apache.flink.kinesis.shaded.io.netty.util.concurrent.SingleThreadEventExecutor$4.run(SingleThreadEventExecutor.java:986)
at org.apache.flink.kinesis.shaded.io.netty.util.internal.ThreadExecutorMap$2.run(ThreadExecutorMap.java:74)
at java.base/java.lang.Thread.run(Thread.java:829)
Caused by: java.io.IOException: Response had content-length of 31 bytes, but only received 0 bytes before the connection was closed.
at org.apache.flink.kinesis.shaded.software.amazon.awssdk.http.nio.netty.internal.ResponseHandler.validateResponseContentLength(ResponseHandler.java:163)
at org.apache.flink.kinesis.shaded.software.amazon.awssdk.http.nio.netty.internal.ResponseHandler.access$700(ResponseHandler.java:75)
at org.apache.flink.kinesis.shaded.software.amazon.awssdk.http.nio.netty.internal.ResponseHandler$PublisherAdapter$1.onComplete(ResponseHandler.java:369)
at org.apache.flink.kinesis.shaded.software.amazon.awssdk.http.nio.netty.internal.nrs.HandlerPublisher.complete(HandlerPublisher.java:447)
at org.apache.flink.kinesis.shaded.software.amazon.awssdk.http.nio.netty.internal.nrs.HandlerPublisher.channelInactive(HandlerPublisher.java:430)
at org.apache.flink.kinesis.shaded.io.netty.channel.AbstractChannelHandlerContext.invokeChannelInactive(AbstractChannelHandlerContext.java:262)
at org.apache.flink.kinesis.shaded.io.netty.channel.AbstractChannelHandlerContext.invokeChannelInactive(AbstractChannelHandlerContext.java:248)
at org.apache.flink.kinesis.shaded.io.netty.channel.AbstractChannelHandlerContext.fireChannelInactive(AbstractChannelHandlerContext.java:241)
at org.apache.flink.kinesis.shaded.io.netty.channel.ChannelInboundHandlerAdapter.channelInactive(ChannelInboundHandlerAdapter.java:81)
at org.apache.flink.kinesis.shaded.io.netty.handler.timeout.IdleStateHandler.channelInactive(IdleStateHandler.java:277)
at org.apache.flink.kinesis.shaded.io.netty.channel.AbstractChannelHandlerContext.invokeChannelInactive(AbstractChannelHandlerContext.java:262)
at org.apache.flink.kinesis.shaded.io.netty.channel.AbstractChannelHandlerContext.invokeChannelInactive(AbstractChannelHandlerContext.java:248)
at org.apache.flink.kinesis.shaded.io.netty.channel.AbstractChannelHandlerContext.fireChannelInactive(AbstractChannelHandlerContext.java:241)
at org.apache.flink.kinesis.shaded.io.netty.channel.DefaultChannelPipeline$HeadContext.channelInactive(DefaultChannelPipeline.java:1405)
at org.apache.flink.kinesis.shaded.io.netty.channel.AbstractChannelHandlerContext.invokeChannelInactive(AbstractChannelHandlerContext.java:262)
at org.apache.flink.kinesis.shaded.io.netty.channel.AbstractChannelHandlerContext.invokeChannelInactive(AbstractChannelHandlerContext.java:248)
at org.apache.flink.kinesis.shaded.io.netty.channel.DefaultChannelPipeline.fireChannelInactive(DefaultChannelPipeline.java:901)
at org.apache.flink.kinesis.shaded.io.netty.handler.codec.http2.AbstractHttp2StreamChannel$Http2ChannelUnsafe$2.run(AbstractHttp2StreamChannel.java:737)
... 6 more 

Related AWS SDK issue: https://github.com/aws/aws-sdk-java-v2/issues/3335

AWS SDK fix: https://github.com/aws/aws-sdk-java-v2/pull/3855/files

 

This mishandled exception creates a deadlock situation that prevents the connectors from making any progress.

We should update the AWS SDK v2 to 2.20.32: https://github.com/aws/aws-sdk-java-v2/commit/eb5619e24e4eaca6f80effa1c43c0cd409cdd53e

Issue Links

• links to: GitHub Pull Request #66
• links to: GitHub Pull Request #67
• links to: GitHub Pull Request #68