-
Bug
-
Resolution: Cannot Reproduce
-
Major
Agents (swarm-client in use) go offline during job execution
00:00:00.002 [EnvInject] - Loading node environment variables.
00:08:49.244 FATAL: java.nio.channels.ClosedChannelException
00:08:49.249 java.nio.channels.ClosedChannelException
00:08:49.249 Also: hudson.remoting.Channel$CallSiteStackTrace: Remote call to JNLP4-connect connection from ************/************
00:08:49.249 at hudson.remoting.Channel.attachCallSiteStackTrace(Channel.java:1737)
00:08:49.249 at hudson.remoting.Request.call(Request.java:197)
00:08:49.249 at hudson.remoting.Channel.call(Channel.java:951)
00:08:49.249 at hudson.FilePath.act(FilePath.java:1104)
00:08:49.249 at org.jenkinsci.plugins.envinject.service.EnvironmentVariablesNodeLoader.gatherEnvVarsForNode(EnvironmentVariablesNodeLoader.java:64)
00:08:49.249 at org.jenkinsci.plugins.envinject.EnvInjectListener.loadEnvironmentVariablesNode(EnvInjectListener.java:80)
00:08:49.249 at org.jenkinsci.plugins.envinject.EnvInjectListener.setUpEnvironment(EnvInjectListener.java:44)
00:08:49.249 at hudson.model.AbstractBuild$AbstractBuildExecution.createLauncher(AbstractBuild.java:542)
00:08:49.249 at hudson.model.AbstractBuild$AbstractBuildExecution.run(AbstractBuild.java:462)
00:08:49.249 at hudson.model.Run.execute(Run.java:1727)
00:08:49.249 at hudson.model.FreeStyleBuild.run(FreeStyleBuild.java:43)
00:08:49.249 at hudson.model.ResourceController.execute(ResourceController.java:97)
00:08:49.249 at hudson.model.Executor.run(Executor.java:429)
00:08:49.249 Caused: hudson.remoting.RequestAbortedException
00:08:49.249 at hudson.remoting.Request.abort(Request.java:335)
00:08:49.249 at hudson.remoting.Channel.terminate(Channel.java:1034)
00:08:49.249 at org.jenkinsci.remoting.protocol.impl.ChannelApplicationLayer.onReadClosed(ChannelApplicationLayer.java:208)
00:08:49.249 at org.jenkinsci.remoting.protocol.ApplicationLayer.onRecvClosed(ApplicationLayer.java:222)
00:08:49.249 at org.jenkinsci.remoting.protocol.ProtocolStack$Ptr.onRecvClosed(ProtocolStack.java:832)
00:08:49.249 at org.jenkinsci.remoting.protocol.FilterLayer.onRecvClosed(FilterLayer.java:287)
00:08:49.250 at org.jenkinsci.remoting.protocol.impl.SSLEngineFilterLayer.onRecvClosed(SSLEngineFilterLayer.java:172)
00:08:49.250 at org.jenkinsci.remoting.protocol.ProtocolStack$Ptr.onRecvClosed(ProtocolStack.java:832)
00:08:49.250 at org.jenkinsci.remoting.protocol.NetworkLayer.onRecvClosed(NetworkLayer.java:154)
00:08:49.250 at org.jenkinsci.remoting.protocol.impl.NIONetworkLayer.ready(NIONetworkLayer.java:179)
00:08:49.250 at org.jenkinsci.remoting.protocol.IOHub$OnReady.run(IOHub.java:789)
00:08:49.250 at jenkins.util.ContextResettingExecutorService$1.run(ContextResettingExecutorService.java:28)
00:08:49.250 at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149)
00:08:49.250 at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624)
00:08:49.250 at java.lang.Thread.run(Thread.java:748)
00:08:49.254 ERROR: Step ?Delete workspace when build is done? failed: no workspace for docker/***************-build #30
00:08:49.254 Finished: FAILURE
Java version: 8
Jenkins and build agents running in docker
[JENKINS-57010] Build agent offline with "FATAL: java.nio.channels.ClosedChannelException"
this seems to be a common characteristic:
00:00:00.002 [EnvInject] - Loading node environment variables.
Jeff Thompson
added a comment - ewel, could you clarify – is this a log from the agent from the master?
Jeff Thompson
added a comment - ewel , could you clarify – is this a log from the agent from the master? Jeff Thompson
added a comment - Since there has been a lack of response for quite a while on this, I'm going to close this issue. If you have additional information to help move this along, feel free to provide it and re-open as needed.
Jeff Thompson
added a comment - Since there has been a lack of response for quite a while on this, I'm going to close this issue. If you have additional information to help move this along, feel free to provide it and re-open as needed. 
ewel, does it commonly fail at that some point, with `EnvironmentVariablesNodeLoader`? Or something similar?
Generally a `ClosedChannelException` is caused by system, network, or environmental issues. Something outside the agent / remoting connection causes the TCP connection to drop. The messages in Jenkins indicate that has happened but do not have any information about the external cause of the disconnection. A heavily loaded master or other resources can cause network connections to drop. I recommend investigating whether the failures share any common characteristics, including stack trace, system type, load, or anything else.