#2 Execution Order & DAGs

Step 01 - Execution Order

The steps listed here assume you are using the new navigation.
To switch, click on your image/avatar in the header, and scroll down to Navigation redesign, where you can disable/enable the New navigation.

  1. If you are coming right from the last track you should still be on the Pipelines page. If you navigated away, just go back to your project (Workshop Project) and use the left hand navigation menu to click through Build -> Pipelines and click the hyperlink starting with # of the most recently run pipeline.

Example: #577244133

  1. You can see that our pipeline jobs ran sequentially, but what if we wanted 2 jobs to run in parallel? We can do that with the needs keyword. Lets navigate back to our .gitlab-ci.yml file to accomplish this.

  2. To edit our pipeline use the left hand navigation menu to click through Build -> Pipeline editor. Right now we only have the unit_test job running during the test stage, so let's add the code_quality job below the unit_test job.:

    code_quality:
  stage: test
  script:
    -  echo "This will run code quality in the future"
  ### This job can run independently of any previous job completion, i.e. build job, in order to execute
  needs: []

  3. Now that we have the two jobs we also want to modify the execution order so that they run at the same time. Add the following line to the end of the unit_test code block:

    needs: []

    The new unit_test should look like this:

    unit_test:
  stage: test
  before_script:
    - yarn add chai@4.3.7 mocha mocha-simple-html-reporter mocha-junit-reporter chai-http@4.3.0 mocha-test-url
  script:
    - ./node_modules/mocha/bin/_mocha "test/*.js" --reporter mocha-junit-reporter --reporter-options mochaFile=./testresults/test-results.xml
    - ./node_modules/mocha/bin/_mocha "test/*.js" --reporter mocha-simple-html-reporter --reporter-options output=./testresults/test-results.html
  after_script:
    - echo "build_app job has run!"
  needs: []

  4. Go ahead and click Commit changes, then use the left hand navigation menu to click through Build -> Pipelines and click the hyperlink that starts with # on the most recently kicked off pipeline. As you watch you can see that the two jobs run in parallel.

If you run into any issues you can use the left hand navigation menu to click through Build -> Pipelines, click Run pipeline, select execution-order and click Run pipeline once again.

Check out relevant docs for needs keyword here

Step 02 - DAG (Directed Acyclic Graph)

  1. Now what if we had many stages and relationships between jobs that we wanted to run as soon as possible? We can accomplish this using a DAG. Use the left hand navigation menu to click through Build -> Pipeline editor so we can create one.

  2. We won't be completing a fully flushed out DAG for this workshop as it is out of scope so we will need some dummy jobs to create what we need. First under the stages section we will want to add a deploy stage so your code should look like this:

    stages:
  - build
  - test
  - deploy

  3. Below are the jobs we want to add under our code_quality job, there are quite a few so it's better to just copy paste the code:

    build_a:
 stage: build
 script:
  - echo "build_a"

build_b:
 stage: build
 script:
  - echo "build_b"

build_c:
 stage: build
 script:
  - echo "build_c"

build_d:
 stage: build
 script:
  - echo "build_d"

build_e:
 stage: build
 script:
  - echo "build_e"

build_f:
 stage: build
 script:
  - echo "build_f"

test_a:
 stage: test
 needs: [build_a]
 script:
  - echo "test_a"

test_b:
 stage: test
 needs: [build_b]
 script:
  - echo "test_b"

test_c:
 stage: test
 needs: [build_c]
 script:
  - echo "test_c"

test_d:
 stage: test
 needs: [build_d]
 script:
  - echo "test_d"

test_e:
 stage: test
 needs: [build_e]
 script:
  - echo "test_e"

test_f:
 stage: test
 needs: [build_f]
 script:
  - echo "test_f"

deploy_a:
 stage: deploy
 needs: [test_a]
 script:
  - echo "deploy_a"

deploy_b:
 stage: deploy
 needs: [test_b]
 script:
  - echo "deploy_b"

deploy_c:
 stage: deploy
 needs: [test_c]
 script:
  - echo "deploy_c"

deploy_d:
 stage: deploy
 needs: [test_d]
 script:
  - echo "deploy_d"

deploy_e:
 stage: deploy
 needs: [test_e]
 script:
  - echo "deploy_e"

deploy_f:
 stage: deploy
 needs: [test_f]
 script:
  - echo "deploy_f"

  4. If you now click the Visualize tab you can see just how complex the many stages are. Lets go ahead and go back to the Edit tab and click Commit changes.

Check out relevant docs for DAGs here

  1. Once committed use the left hand navigation menu to click through Build -> Pipelines and click the hyperlink starting with # on the most recently kicked off pipeline. Here we can watch all of the paths of our DAG run at the same time.

Example: #577244133

If you run into any issues you can use the left hand navigation menu to click through Build -> Pipelines, click Run pipeline, select DAG and click Run pipeline once again.

Edited by Julie Byrne