Issue: I have my Jenkins master and slave setup on AWS instance. I am build i have a maven deploy task to upload a large artifact to nexus. During which the upload is getting timed out and i am getting “Slave went offline during the build” error. How do i solve this issue.
This issue is normally caused by kernel bug in you slave instances. There are many cases in ec2 instances like this. There could be other reasons as well. Here we will discussing a kernel issue that causes this problem.
The Problem
Normally this issue is caused when the TX (packets going outside) are dropped.
Run the following command on the slave machine.
|
1 |
dmesg | grep rocket |
If you get the following output, you can confirm that it’s a kernel bug.
|
1 |
xen_netfront: xennet: skb rides the rocket: 20 slots |
The Solution
You can solve this issue by dropping all the packet segmentation. Execute the following command on your slave machine
|
1 |
sudo ethtool -K eth0 sg off |
The above command should solve the issue. Let us know in the comment section, if your issue has been solved or if there is some other solution to it. Your feedback might help others.