I had the same issue.

Downgrading the Jenkins master kernel from 3.14.19-17.43.amzn1.x86_64 to 3.4.62-53.42.amzn1.x86_64 solved it.

Same thing for us: Amazon Linux master with EC2 slaves plugin and Amazon Linux slaves. Builds were randomly hanging and slaves were timing out.
We downgraded the kernel on the master this morning from 3.14.23-22.44.amzn1.x86_64 to 3.4.73-64.112.amzn1.x86_64 and haven't seen any issues since.
Slaves are still running 3.14 kernel (3.14.27-25.47.amzn1.x86_64).
I'll report back later.

I'm on Amazon EC2 w/ Ubuntu 14.04.1 LTS.

I've downgraded the kernel (manually) to 3.4.

wget http://kernel.ubuntu.com/~kernel-ppa/mainline/v3.4.105-quantal/linux-headers-3.4.105-0304105-generic_3.4.105-0304105.201412012335_amd64.deb
wget http://kernel.ubuntu.com/~kernel-ppa/mainline/v3.4.105-quantal/linux-headers-3.4.105-0304105_3.4.105-0304105.201412012335_all.deb
wget http://kernel.ubuntu.com/~kernel-ppa/mainline/v3.4.105-quantal/linux-image-3.4.105-0304105-generic_3.4.105-0304105.201412012335_amd64.deb
dpkg -i linux-*

Once I had the 3.4 kernel installed, I had to make it the default kernel on boot, so I followed the instructions here: http://statusq.org/archives/2012/10/24/4584/

I'm trying out some builds now to see how Jenkins behaves...

Downgrading the kernel on the master has definitely fixed it for us.
Running for a week now without any trouble.

I have the same issue. kernel is 3.14. Part of the problem is that when the slave.jar file fails, it does NOT cause the agent to display as offline for the master, so the master keeps trying to send jobs:

Expanded the channel window size to 4MB
[05/11/15 18:02:23] [SSH] Starting slave process: cd "/var/lib/jenkins" && java -Dfile.encoding=UTF8 -jar slave.jar

<===[JENKINS REMOTING CAPACITY]===>ERROR: Unexpected error in launching a slave. This is probably a bug in Jenkins.
java.lang.IllegalStateException: Already connected
at hudson.slaves.SlaveComputer.setChannel(SlaveComputer.java:448)
at hudson.slaves.SlaveComputer.setChannel(SlaveComputer.java:366)
at hudson.plugins.sshslaves.SSHLauncher.startSlave(SSHLauncher.java:945)
at hudson.plugins.sshslaves.SSHLauncher.access$400(SSHLauncher.java:133)
at hudson.plugins.sshslaves.SSHLauncher$2.call(SSHLauncher.java:711)
at hudson.plugins.sshslaves.SSHLauncher$2.call(SSHLauncher.java:696)
at java.util.concurrent.FutureTask.run(FutureTask.java:262)
at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145)
at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615)
at java.lang.Thread.run(Thread.java:745)
[05/11/15 18:02:24] Launch failed - cleaning up connection
[05/11/15 18:02:24] [SSH] Connection closed.

Even though the log reports the connection closed, jenkins still reports it as up.

My jenkins master is on 1.596.2.

I was able to connect to the same slave from another jenkins master using the same kernel and jenkins version with no issues...

Hi,

I just do a test on last AWS Linux machine (kernel : 3.14.35), with the same machine on both master and slave.
And the problem gone ...
Jenkins version used is the last stable : 1.609.1

Regards

Related to JENKINS-1948 perhaps?

In AWS, we are using Ubuntu 14.04.4 LTS. EC2-plugin version is 1.36. We are also seeing similar errors where the agent would disconnect from Jenkins randomly with the error below.

ERROR: SEVERE ERROR occurs
org.jenkinsci.lib.envinject.EnvInjectException: hudson.remoting.ChannelClosedException: channel is already closed
at org.jenkinsci.plugins.envinject.service.EnvironmentVariablesNodeLoader.gatherEnvironmentVariablesNode(EnvironmentVariablesNodeLoader.java:79)
at org.jenkinsci.plugins.envinject.EnvInjectListener.loadEnvironmentVariablesNode(EnvInjectListener.java:80)
at org.jenkinsci.plugins.envinject.EnvInjectListener.setUpEnvironment(EnvInjectListener.java:42)
at hudson.model.AbstractBuild$AbstractBuildExecution.createLauncher(AbstractBuild.java:572)
at hudson.model.AbstractBuild$AbstractBuildExecution.run(AbstractBuild.java:492)
at hudson.model.Run.execute(Run.java:1741)
at hudson.model.FreeStyleBuild.run(FreeStyleBuild.java:43)
at hudson.model.ResourceController.execute(ResourceController.java:98)
at hudson.model.Executor.run(Executor.java:410)
Caused by: hudson.remoting.ChannelClosedException: channel is already closed
at hudson.remoting.Channel.send(Channel.java:578)
at hudson.remoting.Request.call(Request.java:130)
at hudson.remoting.Channel.call(Channel.java:780)
at hudson.FilePath.act(FilePath.java:1102)
at org.jenkinsci.plugins.envinject.service.EnvironmentVariablesNodeLoader.gatherEnvironmentVariablesNode(EnvironmentVariablesNodeLoader.java:48)
... 8 more
Caused by: java.io.IOException
at hudson.remoting.Channel.close(Channel.java:1163)
at hudson.slaves.ChannelPinger$1.onDead(ChannelPinger.java:121)
at hudson.remoting.PingThread.ping(PingThread.java:130)
at hudson.remoting.PingThread.run(PingThread.java:86)
Caused by: java.util.concurrent.TimeoutException: Ping started at 1493347954228 hasn't completed by 1493348194229

because there is not recent info here and seems similar to JENKINS-53810 I will close it.