Grouping - Confusion over generated results on dashboard

[parky128]

I have just started using sorry-cypress and I am encountering an issue where it appears I am getting the same specs run on each node.

For example I have a couple of nodes being used for our test stage in our Jenkins Pipeline

stage('Cypress 1') {
        agent { label 'aws_slave_cypress' }
        steps {
            runE2eTests()
        }
    }
    stage('Cypress 2') {
        agent { label 'aws_slave_cypress' }
        steps {
            runE2eTests()
        }
    }
}

This in turn executes the following on each node

sh "CYPRESS_API_URL=\"${CYPRESS_API_URL}\" npx cy2 run --record --key XXX --group 2x-chrome --parallel --ci-build-id ${env.BUILD_ID} --browser chrome --headless --config-file cypress/cypress.${E2E_ENVIRONMENT}.json --env $envCredentials || true"

I have actually only got one spec file in the project current, just whilst I get sorry-cypress setup in our environment.

But I see in the dashboard the same spec gets run twice, one for each node, I can see different machine id's when I click into each of these runs.

Why are these not getting grouped together, what am I missing?

[agoldis]

Please make sure you're using the same buildId If you want two invocation to be within the same run.
Another possible issue is missing indexes on MongoDB "runs" collection - they might disappear in some edge cases and I still didn't have chance to fix this

[parky128]

Ok, so I have used the demo url you supplied ion your previous comment and have performed a couple of runs on our Jenkins, one with and one without the group argument.

See runs here https://sorry-cypress-demo.herokuapp.com/o3-portero/runs

I have two nodes being used during the build job and only one spec file in total.

So I'm seeing the same behaviour with our own sorry-cypress instance. If you can look into the logs and tell me what Im doing wrong configuration wise then I will be forever in your debt!

Thanks

[agoldis]

@parky128
Thanks for generating those runs.

They have different git commits thus being recognized as different runs:

sorry-cypress/packages/director/src/lib/hash.ts

Line 10 in 0ec6ea7

return md5(ciBuildId + sha + projectId);

• https://sorry-cypress-demo.herokuapp.com/run/3516029de03353214c95ab72578318fe points to sha 754c72bd401cf67f327449640a0dffee5338d02e
• https://sorry-cypress-demo.herokuapp.com/run/5907d8ad59b91e21739fed089c11b533 points to sha efb52dd25391294f0d0074522667759696995e89

[parky128]

Oh right, well that's fried my brain! I'll go figure out why that's the case, but thanks for the explanation

[parky128]

Hi @agoldis

I am just coming back to this after some vacation, and looking closer at why different commits are being generated for each run.

I am using Jenkins and have a pipeline setup with a parallel stage configured to spin up a couple of AWS EC2 instances and run the cypress test commands against. I can see in the generated Jenkins build output that for each parallel stage, its cloning my repo and creating independent merge commits which is why they are getting treated as separate runs.

Not sure how familiar you are with Jenkins pipelines, but I have this setup right now as a stage:

stage('E2E-PR') {
    parallel {
        stage('Cypress Tests 1') {
            agent { label 'aws_slave_cypress' }
            steps {
                runE2eTests()
            }
        }
        stage('Cypress Tests 2') {
            agent { label 'aws_slave_cypress' }
            steps {
                runE2eTests()
            }
        }
    }
}

I have not yet figured out how to get around this, one thing I am trying out is getting the parallel runs to use the same original generated workspace from the start of my pipeline on my master (parent) node.

Do you have any Jenkins experience and perhaps some insight into what I could\should do here?

Thanks

[parky128]

Ok, ignore my previous message above, I am now seeing the runs grouped together, no longer getting the different commit ids causing me the problem. I think it was something to do with my slave nodes getting messed up with my running all manner of experiments on them, i just nuked them and got new instances spun up on my Jenkins, and now its all behaving.