-
Bug
-
Resolution: Fixed
-
Minor
-
None
-
Jenkins Instance (2.135) with 5 agents all connected using Kafka (no security)
remoting Kafka plugin 1.1, built with remoting 3.23
Using Remoting Kafka Plugin to version 1.1 (changed the pom file to use the remoting.version=3.23 to match the version on my set up)
I've set a Jenkins instance with 5 nodes all connected with kafka. It did run jobs for about 3 days with a considerable load without failures. But today I had the following failure (full log attached):
Error when executing always post condition:hudson.remoting.ChannelClosedException: Channel "unknown": Remote call on daybreakk failed. The channel is closing down or has closed down
The process that I started in the kafka agent with:
> java -jar remoting-kafka-agent.jar -name daybreakk -master http://jenkins:8081/ -secret 6e6c09c25550bd721cef2efba02bfbd54d6306fc711c2482450abb5ee454cd7f -kafkaURL kafka.al.com.au:9093 -noauth
exited without any log.
When I restarted the process, after all the kafka initial logs it printed: (full log attached)
Sep 04, 2018 8:44:47 AM hudson.remoting.ExportTable unexportByOidSEVERE: Trying to unexport an object that's already unexportedjava.util.concurrent.ExecutionException: Invalid object ID 750 iota=2 at hudson.remoting.ExportTable.diagnoseInvalidObjectId(ExportTable.java:478) at hudson.remoting.ExportTable.unexportByOid(ExportTable.java:516) at hudson.remoting.Channel.unexport(Channel.java:803) at hudson.remoting.Channel.unexport(Channel.java:793) at hudson.remoting.UnexportCommand.execute(UnexportCommand.java:49) at hudson.remoting.Channel$1.handle(Channel.java:565) at hudson.remoting.SynchronousCommandTransport$ReaderThread.run(SynchronousCommandTransport.java:87) Sep 04, 2018 8:44:47 AM hudson.remoting.ExportTable unexportByOidSEVERE: 2nd unexport attempt is hereCommand Unexport created at at hudson.remoting.Command.<init>(Command.java:79) at hudson.remoting.Command.<init>(Command.java:62) at hudson.remoting.UnexportCommand.<init>(UnexportCommand.java:35) at hudson.remoting.RemoteInvocationHandler$PhantomReferenceImpl.cleanup(RemoteInvocationHandler.java:393) at hudson.remoting.RemoteInvocationHandler$PhantomReferenceImpl.access$1000(RemoteInvocationHandler.java:352) at hudson.remoting.RemoteInvocationHandler$Unexporter.run(RemoteInvocationHandler.java:610) at java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:511) at java.util.concurrent.FutureTask.run(FutureTask.java:266) at hudson.remoting.AtmostOneThreadExecutor$Worker.run(AtmostOneThreadExecutor.java:112) at java.lang.Thread.run(Thread.java:745)Caused by: java.lang.Exception: Proxy hudson.remoting.RemoteInvocationHandler@2ee was created for interface hudson.plugins.git.IGitAPI at hudson.remoting.RemoteInvocationHandler.<init>(RemoteInvocationHandler.java:142) at hudson.remoting.RemoteInvocationHandler.wrap(RemoteInvocationHandler.java:161) at hudson.remoting.Channel.export(Channel.java:768) at hudson.remoting.Channel.export(Channel.java:731) at org.jenkinsci.plugins.gitclient.LegacyCompatibleGitAPIImpl.writeReplace(LegacyCompatibleGitAPIImpl.java:198) at sun.reflect.GeneratedMethodAccessor64.invoke(Unknown Source) at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43) at java.lang.reflect.Method.invoke(Method.java:498) at java.io.ObjectStreamClass.invokeWriteReplace(ObjectStreamClass.java:1218) at java.io.ObjectOutputStream.writeObject0(ObjectOutputStream.java:1136) at java.io.ObjectOutputStream.writeObject(ObjectOutputStream.java:348) at hudson.remoting.UserRequest._serialize(UserRequest.java:264) at hudson.remoting.UserRequest.serialize(UserRequest.java:273) at hudson.remoting.UserRequest.perform(UserRequest.java:223) at hudson.remoting.UserRequest.perform(UserRequest.java:54) at hudson.remoting.Request$2.run(Request.java:369) at hudson.remoting.InterceptingExecutorService$1.call(InterceptingExecutorService.java:72) at java.util.concurrent.FutureTask.run(FutureTask.java:266) at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149) at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624) at io.jenkins.plugins.remotingkafka.Engine$1.lambda$newThread$0(Engine.java:47) at java.lang.Thread.run(Thread.java:748) ... ... ... Sep 04, 2018 8:44:47 AM hudson.remoting.Request$2 runINFO: Failed to send back a reply to the request hudson.remoting.Request$2@7da8bc0d: hudson.remoting.ChannelClosedException: Channel "hudson.remoting.Channel@51a38854:daybreakk": channel is already closed[Thread-2] INFO org.apache.kafka.common.utils.AppInfoParser - Kafka version : 1.1.0[Thread-2] INFO org.apache.kafka.common.utils.AppInfoParser - Kafka commitId : fdcf75ea326b8e07
it printed this repeatedly and it successfully connected at some point.
There were no meaningful logs in the support folder of the Remote root directory (full logs attached)
My expectation was that even if we had an exception like this, the job will not fail due to this cause, and keep waiting until the node gets reconnected and in the worse case, fail due to a timeout.
[JENKINS-53397] ChannelClosedException when using kafka nodes
Disregard my last comment, I just saw oleg_nenashev comments on the PR
Just wanted to mention that since I build that against remoting 3.26 I haven't seen this issue again
Thanks for trying it fnaum, last time I upgraded remoting 3.26 in my machine it has a strange exception, I will debug it and spin a release to fix these problems soon.
Jeff Thompson
added a comment - That's good news, fnaum.
pvtuan10, I would be surprised at Remoting 3.26 causing a new exception. Let me know if you continue to see issues there.
Jeff Thompson
added a comment - That's good news, fnaum .
pvtuan10 , I would be surprised at Remoting 3.26 causing a new exception. Let me know if you continue to see issues there. Hi jthompson, Im getting this error log when connecting master with kafka agent in remoting 3.26, which doesn't happen when I use 3.25.
java.util.concurrent.CancellationExceptionjava.util.concurrent.CancellationException at java.util.concurrent.FutureTask.report(FutureTask.java:121) at java.util.concurrent.FutureTask.get(FutureTask.java:192) at io.jenkins.plugins.remotingkafka.KafkaComputerLauncher.launch(KafkaComputerLauncher.java:119) at hudson.slaves.SlaveComputer$1.call(SlaveComputer.java:294) at jenkins.util.ContextResettingExecutorService$2.call(ContextResettingExecutorService.java:46) at jenkins.security.ImpersonatingExecutorService$2.call(ImpersonatingExecutorService.java:71) at java.util.concurrent.FutureTask.run(FutureTask.java:266) at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149) at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624) at java.lang.Thread.run(Thread.java:748) Oct 02, 2018 3:28:58 PM hudson.slaves.ChannelPinger installSEVERE: Failed to set up a ping for testjava.lang.InterruptedException at java.lang.Object.wait(Native Method) at hudson.remoting.Request.call(Request.java:177) at hudson.remoting.Channel.call(Channel.java:954) at hudson.slaves.ChannelPinger.install(ChannelPinger.java:112) at hudson.slaves.ChannelPinger.preOnline(ChannelPinger.java:95) at hudson.slaves.SlaveComputer.setChannel(SlaveComputer.java:660) at hudson.slaves.SlaveComputer.setChannel(SlaveComputer.java:465) at io.jenkins.plugins.remotingkafka.KafkaComputerLauncher$1.call(KafkaComputerLauncher.java:88) at io.jenkins.plugins.remotingkafka.KafkaComputerLauncher$1.call(KafkaComputerLauncher.java:74) at java.util.concurrent.FutureTask.run(FutureTask.java:266) at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149) at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624) at java.lang.Thread.run(Thread.java:748)
Is there any special changed thing in 3.26 related to this?
And cc oleg_nenashev, I still can connect Kafka agent to master and do the build when having this exception, seem like another plugin issue 
Released version 1.1.1 with remoting 3.26, this should not happen again.
Jeff Thompson
added a comment - pvtuan10 I have no idea why you would see that stack trace. There shouldn't be anything in 3.26 that would cause it. But, there was another long-standing issue that showed up for someone in 3.26 so something may have tweaked some timing in Remoting.
Looks like you must have gotten it figured out well enough to proceed, though. Great!
Jeff Thompson
added a comment - pvtuan10 I have no idea why you would see that stack trace. There shouldn't be anything in 3.26 that would cause it. But, there was another long-standing issue that showed up for someone in 3.26 so something may have tweaked some timing in Remoting.
Looks like you must have gotten it figured out well enough to proceed, though. Great! Hey pvtuan10, We have 2 different instances of Jenkins in production. Currently, I'm using it on the most contained, lower risk instance, only 5 nodes and 2 dozens of jobs.
I'm planning to move the other instance that has 12 nodes and 400 jobs soon.
Hi pvtuan10, thanks for the quick turn around.
I do not know much about jenkins/remoting dependencies, but I noticed in your PR that you have jenkins.version=2.129 , and the remoting version 3.26 is only bundled on Jenkins 2.141 (https://jenkins.io/changelog/) Shouldn't this need to be updated as well?
F