IBM i CI failures

[richardlau]

It looks like the IBM CI has been red for as long as we've got history: https://ci.nodejs.org/job/node-test-commit-ibmi/

From the latest CI run, https://ci.nodejs.org/job/node-test-commit-ibmi/798/#showFailuresLink, filtering out the tests already marked flaky we have these failures:

• test.parallel/test-fs-readSync-position-validation test: allow EOVERFLOW errors in fs position tests #43510
• test.parallel/test-icu-env
• test.parallel/test-fs-read-position-validation test: allow EOVERFLOW errors in fs position tests #43510
• test.parallel/test-v8-serialize-leak test: skip test-v8-serialize-leak on IBM i #43511

---

test.parallel/test-fs-readSync-position-validation

node:fs:747
  handleErrorFromBinding(ctx);
  ^

Error: EOVERFLOW: value too large for defined data type, read
    at Object.readSync (node:fs:747:3)
    at testValid (file:///home/IOJS/build/workspace/node-test-commit-ibmi/nodes/ibmi73-ppc64/test/parallel/test-fs-readSync-position-validation.mjs:20:8)
    at file:///home/IOJS/build/workspace/node-test-commit-ibmi/nodes/ibmi73-ppc64/test/parallel/test-fs-readSync-position-validation.mjs:60:3
    at ModuleJob.run (node:internal/modules/esm/module_job:197:25)
    at async Promise.all (index 0)
    at async ESMLoader.import (node:internal/modules/esm/loader:549:24)
    at async loadESM (node:internal/process/esm_loader:83:5)
    at async handleMainPromise (node:internal/modules/run_main:65:12) {
  errno: -127,
  syscall: 'read',
  code: 'EOVERFLOW'
}

Node.js v19.0.0-pre

test.parallel/test-icu-env

timeout

test.parallel/test-fs-read-position-validation

node:internal/process/esm_loader:89
    internalBinding('errors').triggerUncaughtException(
                              ^

[Error: EOVERFLOW: value too large for defined data type, read] {
  errno: -127,
  code: 'EOVERFLOW',
  syscall: 'read'
}

Node.js v19.0.0-pre

test.parallel/test-v8-serialize-leak

node:assert:399
    throw err;
    ^

AssertionError [ERR_ASSERTION]: before=0 after=0
    at Object.<anonymous> (/home/IOJS/build/workspace/node-test-commit-ibmi/nodes/ibmi73-ppc64/test/parallel/test-v8-serialize-leak.js:21:3)
    at Module._compile (node:internal/modules/cjs/loader:1116:14)
    at Module._extensions..js (node:internal/modules/cjs/loader:1170:10)
    at Module.load (node:internal/modules/cjs/loader:994:32)
    at Module._load (node:internal/modules/cjs/loader:834:12)
    at Function.executeUserEntryPoint [as runMain] (node:internal/modules/run_main:81:12)
    at node:internal/main/run_main_module:17:47 {
  generatedMessage: false,
  code: 'ERR_ASSERTION',
  actual: false,
  expected: true,
  operator: '=='
}

Node.js v19.0.0-pre

cc @nodejs/platform-ibmi

[richardlau]

• test.parallel/test-fs-readSync-position-validation
• test.parallel/test-fs-read-position-validation

These two are variations on #36925 (comment). I'll open a PR to add EOVERFLOW to the allowed list of errors.

[richardlau]

• test.parallel/test-v8-serialize-leak

I suspect this one is that IBM i doesn't (cannot?) report memory usage: #32043 (comment)

[richardlau]

@V-for-Vasili, since you volunteered to look, please take a look at test-icu-env on IBM i. It was added back in May by #42683 and appears to be timing out on IBM i, e.g. https://ci.nodejs.org/job/node-test-commit-ibmi/798/nodes=ibmi73-ppc64/testReport/junit/(root)/test/parallel_test_icu_env/

[V-for-Vasili]

Looks like the whole test takes about 3m to complete on IBMi compared to about 11seconds on x64 linux.
All test assertions take comparable time to complete (there is not one that is significantly longer then the rest).
I think marking it SLOW could be an acceptable solution here.

Does not look like SLOW is an option for Release build, PR to skip: #43559

[richardlau]

I edited the IBM i Jenkins job to increase the test timeout from the default 2 mins to 5 mins (via TEST_CI_ARGS="-t 300").

test-icu-env passed in https://ci.nodejs.org/job/node-test-commit-ibmi/803/nodes=ibmi73-ppc64/#showFailuresLink 🎉

It looks like we've gained two new failures in

• test.parallel/test-repl-persistent-history
• test.parallel/test-repl-programmatic-history

[V-for-Vasili]

I assume this is not needed then #43559 if we intend to keep the timeout 5minutes on ibmi?

[richardlau]

I assume this is not needed then #43559 if we intend to keep the timeout 5minutes on ibmi?

Yeah, sorry about that. I'd be happy to keep the timeout on IBM i at 5 minutes (or even extend further if we find it necessary to do so).