Kube burner install latest inconsistent/failing

[paigerube14]

Describe the bug
Kube burner randomly is not able to get the latest kube-burner version. Would be good to either have a backup version or retry the curl command 2 or 3 times to try to get around this. And if it still fails, throw a better error

08-28 11:05:00.106  + ./run.sh
08-28 11:05:00.106  + tee kube-burner-ocp.out
08-28 11:05:00.106  jq: error (at <stdin>:1): Cannot iterate over null (null)

To Reproduce
Steps to reproduce the behavior:

1. Checkout e2e-benchmark
2. cd workloads/kube-burner-ocp-wrapper
3. /run.sh
4. See error

Expected behavior
The workload runs as expected and no error is thrown

Screenshots
https://mastern-jenkins-csb-openshift-qe.apps.ocp-c1.prod.psi.redhat.com/job/scale-ci/job/e2e-benchmarking-multibranch-pipeline/job/kube-burner-ocp/186/console

Additional context

CC: @nathan-weinberg

[paigerube14]

Are we also able to install jq as a requirement with e2e-benchmarking by default. I had to add pip installs into our jenkins jobs to get around this because the system might not have it by default

[vishnuchalla]

@paigerube14 Seems like its just one failed run, due to some down time we had during our new release was being cut along with the artifacts being generated. The later runs seems to be working fine after 11:13 here. Let us know if you come across the same issue again.

[paigerube14]

I've hit this issue off and on many times (don't have exact record) but spitting out a better error message and retrying 2 or 3 times with a small wait inbetween would be very helpful

[paigerube14]

Another example: https://mastern-jenkins-csb-openshift-qe.apps.ocp-c1.prod.psi.redhat.com/job/scale-ci/job/e2e-benchmarking-multibranch-pipeline/job/kube-burner-ocp/178/console

[qiliRedHat]

I met this issue several times yesterday. Example:
https://mastern-jenkins-csb-openshift-qe.apps.ocp-c1.prod.psi.redhat.com/job/scale-ci/job/e2e-benchmarking-multibranch-pipeline/job/kube-burner-ocp/230/console

[svetsa-rh]

another one https://mastern-jenkins-csb-openshift-qe.apps.ocp-c1.prod.psi.redhat.com/job/scale-ci/job/e2e-benchmarking-multibranch-pipeline/job/kube-burner-ocp/240/console

Not sure if this would help, but here is m-g https://mastern-jenkins-csb-openshift-qe.apps.ocp-c1.prod.psi.redhat.com/job/scale-ci/job/e2e-benchmarking-multibranch-pipeline/job/must-gather/261/

[vishnuchalla]

looks like there is some delay/error in curl command execution (for fetching latest tagged assets) in jenkins test setup. I would recommend to use KUBE_BURNER_VERSION by specifying the desired version of kube-burner to be executed (for example v1.7.6 or so) in the job params for OCP runs like we are using KUBE_BURNER_URL var for non-ocps runs here. This will help us have stable and consistent runs

[rsevilla87]

Can we close this issue @paigerube14 ?

[paigerube14]

Yes!