-
Notifications
You must be signed in to change notification settings - Fork 235
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Split pytest by 'slow_test' tag and run from different k8s pods to reduce premerge job duration #3241
Conversation
build |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
From Jenkins it still looks like the job took 2:17h to complete. From the log it appears that both jobs were running in parallel, so I really would like to understand why this not any faster.
Any change to Jenkinsfile will take effect after merge due to security reason. Please refer to Jenkins job |
… memory Signed-off-by: Alex Zhang <alex4zhang@gmail.com>
…duce premerge job duration and keep one Spark version of the unit test for premerge job Signed-off-by: Alex Zhang <alex4zhang@gmail.com>
build |
Signed-off-by: Alex Zhang <alex4zhang@gmail.com>
build |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
LGTM, thx~
We will need to deal w/ TODO later
TODO: this function should be named as 'integration_test()' but it would break backward compatibility
This commented out 3 of the 4 sets of scala tests that were being run before. I asked about it in this PR and it was not addressed. |
it's already replied in #3241 (comment) . Please let me know if you have any other comments |
…duce premerge job duration (NVIDIA#3241) * Add slow_test tag for pytest files that required long duration or big memory Signed-off-by: Alex Zhang <alex4zhang@gmail.com> * Split pytest by 'slow test' tag and run from different k8s pods to reduce premerge job duration and keep one Spark version of the unit test for premerge job Signed-off-by: Alex Zhang <alex4zhang@gmail.com> * Simplify description of 'slow_test' mark Signed-off-by: Alex Zhang <alex4zhang@gmail.com> Signed-off-by: Raza Jafri <rjafri@nvidia.com>
part of #2731
split pytests by 'slow_test' tag in order to balance test duration and memory consumption from two running pods in parallel after identify two .py files that required long duration or big memory. Now premerge overall duration could be reduced to ~1:30h from ~2:30h