fix potential deadlock in pulsar client close #5731
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
rerun integration tests |
aahmed-se
approved these changes
Nov 25, 2019
sijie
approved these changes
Nov 26, 2019
jerrypeng
added a commit
to jerrypeng/incubator-pulsar
that referenced
this pull request
Dec 17, 2019
# Motivation A deadlock on PulsarClient.close() can happen if there are producers/consumers that are not closed prior to calling PulsarClient.close() and we have to close them in the routine. The deadlock happens be cause we piggy pack off of a "pulsar-io" thread that is used to shutdown the connection in producer/consumer close, to also shutdown the EventLoopGroup. "pulsar-io" thread is part of the EventLoopGroup thus it tries to shutdown itself and causes a deadlock. Below is a stacktrace of what it look like: ``` "pulsar-client-io-1-1" apache#20 prio=5 os_prio=31 tid=0x00007fc312a78800 nid=0x9a03 in Object.wait() [0x000070000384e000] java.lang.Thread.State: WAITING (on object monitor) at java.lang.Object.wait(Native Method) - waiting on <0x00000007b8532640> (a io.netty.util.concurrent.DefaultPromise) at java.lang.Object.wait(Object.java:502) at io.netty.util.concurrent.DefaultPromise.await(DefaultPromise.java:252) - locked <0x00000007b8532640> (a io.netty.util.concurrent.DefaultPromise) at io.netty.util.concurrent.DefaultPromise.sync(DefaultPromise.java:403) at io.netty.util.concurrent.DefaultPromise.sync(DefaultPromise.java:35) at org.apache.pulsar.client.impl.ConnectionPool.close(ConnectionPool.java:286) at org.apache.pulsar.client.impl.PulsarClientImpl.shutdown(PulsarClientImpl.java:578) at org.apache.pulsar.client.impl.PulsarClientImpl.lambda$closeAsync$18(PulsarClientImpl.java:560) at org.apache.pulsar.client.impl.PulsarClientImpl$$Lambda$82/878861517.run(Unknown Source) at java.util.concurrent.CompletableFuture.uniRun(CompletableFuture.java:705) at java.util.concurrent.CompletableFuture$UniRun.tryFire(CompletableFuture.java:687) at java.util.concurrent.CompletableFuture.postComplete(CompletableFuture.java:474) at java.util.concurrent.CompletableFuture.complete(CompletableFuture.java:1962) at org.apache.pulsar.client.impl.ProducerImpl.lambda$closeAsync$9(ProducerImpl.java:735) at org.apache.pulsar.client.impl.ProducerImpl$$Lambda$80/1123226989.apply(Unknown Source) at java.util.concurrent.CompletableFuture.uniHandle(CompletableFuture.java:822) at java.util.concurrent.CompletableFuture$UniHandle.tryFire(CompletableFuture.java:797) at java.util.concurrent.CompletableFuture.postComplete(CompletableFuture.java:474) at java.util.concurrent.CompletableFuture.complete(CompletableFuture.java:1962) at org.apache.pulsar.client.impl.ClientCnx.handleSuccess(ClientCnx.java:406) at org.apache.pulsar.common.protocol.PulsarDecoder.channelRead(PulsarDecoder.java:222) at io.netty.channel.AbstractChannelHandlerContext.invokeChannelRead(AbstractChannelHandlerContext.java:374) at io.netty.channel.AbstractChannelHandlerContext.invokeChannelRead(AbstractChannelHandlerContext.java:360) at io.netty.channel.AbstractChannelHandlerContext.fireChannelRead(AbstractChannelHandlerContext.java:352) at io.netty.handler.codec.ByteToMessageDecoder.fireChannelRead(ByteToMessageDecoder.java:326) at io.netty.handler.codec.ByteToMessageDecoder.channelRead(ByteToMessageDecoder.java:300) at io.netty.channel.AbstractChannelHandlerContext.invokeChannelRead(AbstractChannelHandlerContext.java:374) at io.netty.channel.AbstractChannelHandlerContext.invokeChannelRead(AbstractChannelHandlerContext.java:360) at io.netty.channel.AbstractChannelHandlerContext.fireChannelRead(AbstractChannelHandlerContext.java:352) at io.netty.channel.DefaultChannelPipeline$HeadContext.channelRead(DefaultChannelPipeline.java:1422) at io.netty.channel.AbstractChannelHandlerContext.invokeChannelRead(AbstractChannelHandlerContext.java:374) at io.netty.channel.AbstractChannelHandlerContext.invokeChannelRead(AbstractChannelHandlerContext.java:360) at io.netty.channel.DefaultChannelPipeline.fireChannelRead(DefaultChannelPipeline.java:931) at io.netty.channel.nio.AbstractNioByteChannel$NioByteUnsafe.read(AbstractNioByteChannel.java:163) at io.netty.channel.nio.NioEventLoop.processSelectedKey(NioEventLoop.java:700) at io.netty.channel.nio.NioEventLoop.processSelectedKeysOptimized(NioEventLoop.java:635) at io.netty.channel.nio.NioEventLoop.processSelectedKeys(NioEventLoop.java:552) at io.netty.channel.nio.NioEventLoop.run(NioEventLoop.java:514) at io.netty.util.concurrent.SingleThreadEventExecutor$6.run(SingleThreadEventExecutor.java:1050) at io.netty.util.internal.ThreadExecutorMap$2.run(ThreadExecutorMap.java:74) at io.netty.util.concurrent.FastThreadLocalRunnable.run(FastThreadLocalRunnable.java:30) at java.lang.Thread.run(Thread.java:748) ``` This deadlock went unnoticed for a long time because in the past we never waited for the shutdown of the EventLoopGroup to complete Also apache#5628 does not solve the issue. This PR should supersede that one.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Motivation
A deadlock on PulsarClient.close() can happen if there are producers/consumers that are not closed prior to calling PulsarClient.close() and we have to close them in the routine. The deadlock happens be cause we piggy pack off of a "pulsar-io" thread that is used to shutdown the connection in producer/consumer close, to also shutdown the EventLoopGroup. "pulsar-io" thread is part of the EventLoopGroup thus it tries to shutdown itself and causes a deadlock.
Below is a stacktrace of what it look like:
This deadlock went unnoticed for a long time because in the past we never waited for the shutdown of the EventLoopGroup to complete
Also #5628 does not solve the issue. This PR should supersede that one.