[Bug]: Tests appear to stall quietly with V28

[phawxby]

Version

28.1.1

Steps to reproduce

I'm not really sure how to demonstrate a repro case for this, it's a private repository and i'm fairly confident at this point it's due to the size of our repository. I'm hoping this provides a place for others to also notify if they're experiencing the same.

My gut feeling at this point is it's to do with us approaching the memory limit on the machine, which is obviously a problem, but it should probably output some kind of more helpful message than just stalling out and not doing anything until it times out. I'm assuming this based on the tests completing without coverage so memory limits aren't a problem.

Node 16.10.0 use because of this issue.

Jest 27.4.1 with coverage

Test Suites: 850 passed, 850 total
Tests:       1 skipped, 6707 passed, 6708 total
Snapshots:   1895 passed, 1895 total
Time:        249.302 s
Ran all test suites.
CircleCI received exit code 0

Jest 28.1.1 with coverage.

 PASS   unit  src/common/helpers/tokens/xrx/product/__tests__/ProductSupportPhoneToken.test.ts
 PASS   unit  src/common/helpers/sitemap/__tests__/segment.test.ts (14.47 s)

Too long with no output (exceeded 5m0s): context deadline exceeded

Test step timed out at 9m 18s. Time in the job up until the start of the test step was about 1m. So roughly 5m from the start of the job until it stopped outputting.

Jest 28.1.1 without coverage.
(Ignoring the test failure, it's a known flakey test we need to fix)

Test Suites: 1 failed, 849 passed, 850 total
Tests:       1 failed, 1 skipped, 6691 passed, 6693 total
Snapshots:   1 failed, 1895 passed, 1896 total
Time:        221.092 s
Ran all test suites.

Exited with code exit status 1
CircleCI received exit code 1

Expected behavior

Some kind of error message regarding lack of memory or something.

Actual behavior

Tests appear to stall with no output.

Additional context

No response

Environment

System:
    OS: macOS 12.4
    CPU: (12) x64 Intel(R) Core(TM) i7-9750H CPU @ 2.60GHz
  Binaries:
    Node: 16.10.0 - ~/.nvm/versions/node/v16.10.0/bin/node
    Yarn: 1.22.10 - /usr/local/bin/yarn
    npm: 6.14.6 - ~/Documents/GitHub/brand-engine/node_modules/.bin/npm
  npmPackages:
    jest: ^28.1.1 => 28.1.1

[phawxby]

I've been far more aggressive with my coveragePathIgnorePatterns and I've finally managed to get it to pass on Jest V28 with coverage enabled.

With sharding enabled split over 2 instances you would expect lower memory usage and an even higher chance of passing.

Instance 1

 PASS   unit  src/common/helpers/tokens/xrx/locale/__tests__/SiteContextTokens.test.ts
 PASS   unit  src/features/product/handlers/compare/__tests__/formatSpec.test.ts

Too long with no output (exceeded 5m0s): context deadline exceeded

Instance 2

 PASS   unit  src/common/helpers/sitemap/__tests__/segment.test.ts (25.583 s)
 PASS   unit  src/common/log/__tests__/logBuffer.test.ts (93.319 s)

Too long with no output (exceeded 5m0s): context deadline exceeded

logBuffer is not a test that should take 93 seconds, it's a very basic test suite that should be pretty much instant.

Memory usage is significantly worse with sharding!?

Edit: I'm going to take this way back and split apart the jobs of upgrading jest, improving coverage output isolation, Node 18+, etc

[skovhus]

@phawxby why was this closed? Did you find a solution? 🙏

We observe the same issue of silently hanging tests on a fairly large test suite.

[phawxby]

@phawxby why was this closed? Did you find a solution? 🙏

We observe the same issue of silently hanging tests on a fairly large test suite.

Not exactly but I've spent some time this week trying to isolate what the hell is going on and break it into constituent parts. This is the first bit.
#13054

[github-actions]

This issue has been automatically locked since there has not been any recent activity after it was closed. Please open a new issue for related bugs.
Please note this issue tracker is not a help forum. We recommend using StackOverflow or our discord channel for questions.