Unexpected "timeout exceeded" fails in BigTest unit tests
Description
CSP Request Details
CSP Rejection Details
Potential Workaround
Attachments
blocks
is duplicated by
relates to
Checklist
hideTestRail: Results
Activity
Zak Burke March 9, 2020 at 7:33 PM
I'm closing this for now given that it's effectively become an umbrella for a bunch of separate memory-leak tickets and OOM errors appear to be the proximate cause for BigTest throwing a "timeout exceeded" error.

Cate Boerema March 5, 2020 at 12:32 PM
Hi can you please put some points on this issue so we can count it towards our bugfixing quota?
Zak Burke February 25, 2020 at 7:42 PM
I filed to turn unit tests on in CI for UIIN and assigned it to this team but not this sprint, at least not yet. 20 tests fail when I run them locally.
I haven't dug into with , but I will probably need some assistance from him and/or to get the CI environment and our webpack configs all in sync. It would be great to do work on this this quarter if you think we can get it in.

Cate Boerema February 25, 2020 at 5:30 PM
Thanks, . Should we create a story for turning the unit tests back on in Inventory and put it in this sprint?
What about FOLIO-2433?
Zak Burke February 25, 2020 at 4:17 PM
, it's a long story. This is essentially a follow-up to problems first identified in . It's a bit of an umbrella issue that covers the problems we've had running unit tests on pull requests. We've identified several separate memory leaks that all manifest in the same way, i.e. failing unit tests:
memory leak in
withRenew.js
memory leak in
Pane
,Paneset
memory leak in mousetrap
memory leak in React in dev mode, i.e.
NODE_ENV=development
The first three are resolved and we are once again able to successfully merge PRs in ui-users. I haven't turned unit tests back on for ui-inventory but that would be worthwhile to attempt at this point.
From times to times it happens that a properly written test for a correctly working code fails because
Sometimes it can take up to 5-10 re-runs (with no changes to the tests or tested code) to make the tests pass. I haven't noticed any regularity, any ways to figure out which tests and when are vulnerable to this kind of fails.