cje-crash-course/exercises/10-distributed-builds/instructions.md

Exercise 10

You will start a VM using Vagrant as practice environment to set up a distributed build with 2 agents.

Configuring and Executing Jobs in a Distributed Build

1. Go to "Manage Jenkins" > "Manage Nodes". You should see a single master node.
2. Configure the master node by setting the # of executor value to 0. That will take care of never using the master for job workload.
3. Have other physical or virtual machines ready that can act as agent nodes. Log into the machine as root user and create a new user named jenkins with the commands useradd -d /var/lib/jenkins jenkins and passwd jenkins. From the master node, copy the contents of the id_rsa.key file to the clipboard. Paste the contents of the clipboard to the file /var/lib/jenkins/.ssh/authorized_keys.
4. Add new nodes by clicking "New Nodes". Enter an appropriate name and select the option "Permanent Agent". Use the remote directory /home/jenkins/jenkins_slave and set the # of executors to 2. Enter the host and provide credentials by selecting "SSH Username with private key". To keep things easy select "Non verifying Verification Strategy".
5. Trigger a build. You should see that the build is only executed on the agent and not the master node.
6. Add at least one more agent.
7. Trigger a build. You should see that the build can be executed on any of the agents.
8. Configure one of the agents to only build jobs with a specific label e.g. java. Change the "Usage" field to "Only build jobs with label expressions matching this node".
9. Configure the gradle-initializr job and assign the label java.
10. Trigger a build of the gradle-initializr job. It is only executed by the dedicated agent.
11. Trigger other jobs that do not have the label java assigned to them. They are waiting for an agent that can execute the build.