Slave went offline during the build issue in AWS Jenkins server

[gkri...@infoblox.com]

Hi Team,

I am using AWS Instance as a Jenkins server which has different kind of nodes ( AWS EC2 Instance, Linux and Windows Nodes )

We are using this Jenkins server to execute UT and Functional test to qualify the builds.

We are facing "Slave went offline during the build" issue frequently. ( Note For EC2 instance its less frequency )

Jenkins version : 1.658

OS : Ubuntu 14.04.2 LTS

I need some help on fixing "Slave went offline during the build" issue.

Observed different error messages in log during the slave went offline, In that following 2 messages we observed frequently

1) SEVERE: I/O error in channel followed by Unexpected termination of the channel

2) TcpSlaveAgentListener$ConnectionHandler error followed by Protocol:JNLP3-connect ( This is mostly in Windows machine )

Whenever the Job failed due to slave went offline issue we are seeing either one of the message as above. 

Please refer the following logs for reference.

Note : 1) We have raised Internal IT Ticket to monitor Packet loss of AWS Instance to my Company network and we confirmed that there were no packet loss

       2) We Raised AWS Support ticket to make sure do we need to tune anything specific with respect in AWS Side and confirmed that everything is perfect.

1) We are connecting the nodes on-demand with 1 minute of delay

2) Disabled "Response Time" in preventive node monitoring settings 

But still we are observing this issue.

It would be great if somebody guides me to fix this issue.

Thanks,

Gopal