I'm beginning to think this is a bad test design; if it is indicating a newly introduced problem it almost certainly wasn't from #5139 which it failed on, and it it isn't then even with all the retry logic around this test it can still fail for no reason.
Maybe we need to demote it to an integration test, or eliminate it altogether in favor of something deterministic.
┆Issue is synchronized with this Jira Story
┆Issue Number: TOIL-1663
As shown in https://ucsc-ci.com/databiosphere/toil/-/jobs/81547, the
JobServiceTest.testServiceParallelRecursive
test can hit its 1200 second timeout and fail.I'm beginning to think this is a bad test design; if it is indicating a newly introduced problem it almost certainly wasn't from #5139 which it failed on, and it it isn't then even with all the retry logic around this test it can still fail for no reason.
Maybe we need to demote it to an integration test, or eliminate it altogether in favor of something deterministic.
┆Issue is synchronized with this Jira Story ┆Issue Number: TOIL-1663