Master-agent architecture in Jenkins - Step-by-Step Execution

Process Flow - Master-agent architecture

Master Node

↓

Assign Job

↓

Agent Node

↓

Execute Job

↓

Send Results

↓

Master Node Receives Results

The Jenkins master assigns jobs to agents, which run the jobs and send results back to the master.

Execution Sample

Jenkins

master.launchJob(job1)
agent1.execute(job1)
agent1.sendResult(master)
master.receiveResult(job1)

Master sends a job to agent, agent runs it, then sends back results to master.

Process Table

Step	Action	Node	Job State	Result Sent	Master State
1	Master assigns job1	Master	job1 assigned	No	Waiting for result
2	Agent1 starts job1	Agent1	job1 running	No	Waiting for result
3	Agent1 completes job1	Agent1	job1 done	Yes	Waiting for result
4	Agent1 sends result	Agent1	job1 done	Yes	Result sent
5	Master receives result	Master	job1 done	Yes	Result received
6	Master updates job status	Master	job1 complete	Yes	Job complete

💡 Job1 completed and result received by master, execution ends.

Status Tracker

Variable	Start	After Step 1	After Step 2	After Step 3	After Step 5	Final
job1_state	not assigned	assigned	running	done	done	complete
result_sent	No	No	No	Yes	Yes	Yes
master_state	idle	waiting for result	waiting for result	waiting for result	result received	job complete

Key Moments - 3 Insights

Why does the master wait after assigning the job?

Because the master waits for the agent to finish and send back the job result, as shown in steps 1 to 5 in the execution table.

Can the agent run the job without the master assigning it first?

No, the master must assign the job first (step 1) before the agent can start running it (step 2).

What happens if the agent does not send the result back?

The master remains in 'waiting for result' state and does not mark the job complete, as seen in the master_state variable in the tracker.

Visual Quiz - 3 Questions

Look at the execution table, at which step does the agent start running the job?

A Step 1

B Step 2

C Step 3

D Step 4

💡 Hint

Check the 'Action' and 'Node' columns for when the agent starts the job.

According to the variable tracker, what is the job1_state after step 3?

A assigned

B running

C done

D complete

💡 Hint

Look at the 'job1_state' row under 'After Step 3' in the variable tracker.

If the agent never sends the result, what will the master's state be after step 5?

A waiting for result

B result received

C job complete

D idle

💡 Hint

Refer to the 'master_state' variable in the tracker and the explanation in key moments.

Concept Snapshot

Jenkins Master-Agent Architecture:
- Master assigns jobs to agents.
- Agents run jobs and send results back.
- Master waits for results before marking jobs complete.
- Agents cannot run jobs without master assignment.
- Communication ensures distributed job execution.

Full Transcript

In Jenkins master-agent architecture, the master node assigns jobs to agent nodes. The agents execute the assigned jobs and send the results back to the master. The master waits for these results before marking the job as complete. This flow ensures that jobs are distributed and managed efficiently. The master controls job assignment and collects results, while agents handle the execution workload.

Practice

1. What is the main role of the Jenkins master in the master-agent architecture?

easy

A. To run jobs on different machines

B. To control and schedule jobs

C. To store job logs only

D. To act as a backup server

Solution

1. Step 1: Understand the master role

   The Jenkins master is responsible for controlling and scheduling jobs, meaning it decides when and where jobs run.

2. Step 2: Differentiate from agent role

   Agents actually run the jobs on separate machines, not the master.

3. Final Answer:

   To control and schedule jobs -> Option B

4. Quick Check:

   Master controls jobs = B [OK]

Hint: Master schedules jobs; agents run them [OK]

Common Mistakes:

• Confusing master with agent role
• Thinking master runs jobs directly
• Assuming master only stores logs

2. Which of the following is the correct way to define an agent in a Jenkins pipeline script?

easy

A. run agent

B. master agent

C. agent any

D. node master

Solution

1. Step 1: Recall Jenkins pipeline syntax

   In Jenkins declarative pipeline, agent any means the job can run on any available agent.

2. Step 2: Check other options

   Options like master agent, run agent, and node master are not valid syntax for defining agents.

3. Final Answer:

   agent any -> Option C

4. Quick Check:

   Correct agent syntax = A [OK]

Hint: Use 'agent any' to run on any available agent [OK]

Common Mistakes:

• Using invalid keywords like 'master agent'
• Confusing 'node' with 'agent'
• Writing 'run agent' which is not valid

3. Given this Jenkins pipeline snippet, what will happen?

pipeline {
  agent { label 'linux' }
  stages {
    stage('Build') {
      steps {
        echo 'Building on Linux agent'
      }
    }
  }
}

medium

A. The job runs on any available agent

B. The job runs on the master node

C. The job fails due to syntax error

D. The job runs on an agent labeled 'linux'

Solution

1. Step 1: Analyze the agent label

   The pipeline specifies agent { label 'linux' }, so it will run on an agent with the label 'linux'.

2. Step 2: Understand job execution

   Since the label is specific, the job will not run on any agent or the master unless they have that label.

3. Final Answer:

   The job runs on an agent labeled 'linux' -> Option D

4. Quick Check:

   Agent label 'linux' = C [OK]

Hint: Agent label selects specific machine for job [OK]

Common Mistakes:

• Assuming 'agent any' runs here
• Thinking job runs on master by default
• Believing syntax causes failure

4. You configured an agent but Jenkins shows it as offline. What is the most likely cause?

medium

A. Agent machine is turned off or unreachable

B. Master node is overloaded

C. Pipeline syntax error in Jenkinsfile

D. Agent label does not match any job

Solution

1. Step 1: Understand agent offline status

   An agent shows offline if Jenkins cannot connect to it, often because the machine is off or network unreachable.

2. Step 2: Eliminate other causes

   Master overload or pipeline syntax errors do not cause agent offline status. Label mismatch prevents job assignment but does not make agent offline.

3. Final Answer:

   Agent machine is turned off or unreachable -> Option A

4. Quick Check:

   Offline agent = unreachable machine [OK]

Hint: Offline agent means machine unreachable or off [OK]

Common Mistakes:

• Blaming master overload for offline agent
• Confusing syntax errors with connectivity
• Thinking label mismatch causes offline

5. How can you improve Jenkins performance when running many jobs simultaneously using master-agent architecture?

hard

A. Add more agents to distribute the workload

B. Run all jobs on the master node to avoid delays

C. Disable agents to reduce network traffic

D. Use only one agent with high CPU

Solution

1. Step 1: Understand workload distribution

   Adding more agents allows Jenkins to run jobs in parallel on different machines, improving performance.

2. Step 2: Evaluate other options

   Running all jobs on master overloads it, disabling agents reduces capacity, and one agent limits parallelism.

3. Final Answer:

   Add more agents to distribute the workload -> Option A

4. Quick Check:

   More agents = better parallel job handling [OK]

Hint: More agents mean faster job execution [OK]

Common Mistakes:

• Overloading master with all jobs
• Disabling agents thinking it helps
• Relying on single powerful agent only