mirror of https://github.com/citusdata/citus.git
We have lots of flaky tests in CI and most of these random failures are very hard/impossible to reproduce locally. This adds a job definition to CI that allows adding a temporary job to rerun the same test in CI a lot of times. This will very often reproduce the random failures. If you then try to change the test or code to fix the random failure, you can confirm that it's indeed fixed by using this job. A future improvement to this job would be to run it (or a variant of it) automatically for every newly added test, and maybe even changed tests. This is not implemented in this PR. An example of this job running can be found here: https://app.circleci.com/pipelines/github/citusdata/citus/26682/workflows/a2638385-35bc-443c-badc-7713a8101313 |
||
---|---|---|
.. | ||
config.yml |