You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
To speed up getting the results of tests, it is possible to parallelize their execution by specifying the --parallel.job_count and --parallel.job_index flags as is done in out Buildkite CI script:
returnfmt.Errorf("root: bad scenario name regexp: %w", err)
}
ifmatch {
matched[scenario] =true
anyMatched=true
}
I think this just means that scenarios will be grouped into different parallel jobs, not that it will miss/skip some cases. If we want to ensure a stable order one possible fix (i think) is sorting this array, once it's built:
SUMMARY
To speed up getting the results of tests, it is possible to parallelize their execution by specifying the
--parallel.job_count
and--parallel.job_index
flags as is done in out Buildkite CI script:oasis-core/.buildkite/scripts/test_e2e.sh
Lines 64 to 65 in 6c6e703
However, the oasis-test-runner currently doesn't correctly partition the jobs to different runners.
ISSUE TYPE
ACTUAL RESULTS
For example, running:
resulted in the following two scenarios being run:
Fuller output:
Running the same command again resulted in the following two scenarios being run:
Fuller output:
EXPECTED RESULTS
Running the same command for parallel job execution should always result in the same set of scenarios being run.
The text was updated successfully, but these errors were encountered: