Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Test /parallel/test-fs-stat-bigint fails 3 out of 100 times #24593

Closed
dominikeinkemmer opened this issue Nov 24, 2018 · 24 comments
Closed

Test /parallel/test-fs-stat-bigint fails 3 out of 100 times #24593

dominikeinkemmer opened this issue Nov 24, 2018 · 24 comments
Assignees
Labels
flaky-test Issues and PRs related to the tests with unstable failures on the CI. fs Issues and PRs related to the fs subsystem / file system. macos Issues and PRs related to the macOS platform / OSX.

Comments

@dominikeinkemmer
Copy link
Contributor

dominikeinkemmer commented Nov 24, 2018

Hey, I just ran the test suit and it seems that the tests for /parallel/test-fs-stat-bigint.js fail about 3 times in 100 runs. When running python2 tools/test.py --repeat=100 -J parallel/test-fs-stat-bigint the output is as follows:

=== release test-fs-stat-bigint ===                    
Path: parallel/test-fs-stat-bigint
(node:56622) ExperimentalWarning: The fs.promises API is experimental
assert.js:351
    throw err;
    ^

AssertionError [ERR_ASSERTION]: atimeMs is not a safe integer, difference should < 1.
Number version 1543043785204.1536, BigInt version 1543043785197n
    at verifyStats (/Users/dominik.einkemmer/Documents/Nodefest/node/test/parallel/test-fs-stat-bigint.js:72:7)
    at fs.stat (/Users/dominik.einkemmer/Documents/Nodefest/node/test/parallel/test-fs-stat-bigint.js:109:7)
    at FSReqCallback.oncomplete (fs.js:162:5)
Command: out/Release/node /Users/dominik.einkemmer/Documents/Nodefest/node/test/parallel/test-fs-stat-bigint.js
=== release test-fs-stat-bigint ===                    
Path: parallel/test-fs-stat-bigint
(node:56623) ExperimentalWarning: The fs.promises API is experimental
assert.js:351
    throw err;
    ^

AssertionError [ERR_ASSERTION]: atimeMs is not a safe integer, difference should < 1.
Number version 1543043785206.173, BigInt version 1543043785199n
    at verifyStats (/Users/dominik.einkemmer/Documents/Nodefest/node/test/parallel/test-fs-stat-bigint.js:72:7)
    at fs.stat (/Users/dominik.einkemmer/Documents/Nodefest/node/test/parallel/test-fs-stat-bigint.js:109:7)
    at FSReqCallback.oncomplete (fs.js:162:5)
Command: out/Release/node /Users/dominik.einkemmer/Documents/Nodefest/node/test/parallel/test-fs-stat-bigint.js
=== release test-fs-stat-bigint ===                    
Path: parallel/test-fs-stat-bigint
assert.js:351
    throw err;
    ^

AssertionError [ERR_ASSERTION]: atimeMs is not a safe integer, difference should < 1.
Number version 1543043785397.4092, BigInt version 1543043785396n
    at verifyStats (/Users/dominik.einkemmer/Documents/Nodefest/node/test/parallel/test-fs-stat-bigint.js:72:7)
    at Object.<anonymous> (/Users/dominik.einkemmer/Documents/Nodefest/node/test/parallel/test-fs-stat-bigint.js:84:3)
    at Module._compile (internal/modules/cjs/loader.js:722:30)
    at Object.Module._extensions..js (internal/modules/cjs/loader.js:733:10)
    at Module.load (internal/modules/cjs/loader.js:620:32)
    at tryModuleLoad (internal/modules/cjs/loader.js:560:12)
    at Function.Module._load (internal/modules/cjs/loader.js:552:3)
    at Function.Module.runMain (internal/modules/cjs/loader.js:775:12)
    at startup (internal/bootstrap/node.js:300:19)
    at bootstrapNodeJSCore (internal/bootstrap/node.js:826:3)
Command: out/Release/node /Users/dominik.einkemmer/Documents/Nodefest/node/test/parallel/test-fs-stat-bigint.js
[00:02|% 100|+  97|-   3]: Done

My system spec:

  • Macbook: MacBook Pro (13-inch, 2018)
  • Operating system: MacOS Mojave v10.14.1
  • Processor: 2.7 GHz Intel Core i7
  • Memory: 16 GB 2133 MHz LPDDR3
@addaleax addaleax added flaky-test Issues and PRs related to the tests with unstable failures on the CI. fs Issues and PRs related to the fs subsystem / file system. labels Nov 24, 2018
@Trott
Copy link
Member

Trott commented Nov 24, 2018

See also: #24565

@Trott
Copy link
Member

Trott commented Nov 24, 2018

/ping @joyeecheung

@refack
Copy link
Contributor

refack commented Nov 25, 2018

Possible fix/clarify - #23821

@refack refack added the macos Issues and PRs related to the macOS platform / OSX. label Nov 25, 2018
@bnoordhuis
Copy link
Member

I haven't been able to reproduce but looking at the test and the output OP posted, I suspect there's something updating the atime between the two stat calls. The tests in test-fs-stat-bigint.js follow this pattern:

  1. stat/fstat/lstat either synchronously or asynchronously with { bigint: true }, then
  2. repeat the operation but this time without { bigint: true }

There's a time window between 1 and 2 where an external actor can update the atime. The three failures all have bigint atime < non-bigint atime, probably not a coincidence.

@bnoordhuis
Copy link
Member

Failure on smartos:

13:01:01 not ok 649 parallel/test-fs-stat-bigint
13:01:01   ---
13:01:01   duration_ms: 0.727
13:01:01   severity: fail
13:01:01   exitcode: 1
13:01:01   stack: |-
13:01:01     assert.js:86
13:01:01       throw new AssertionError(obj);
13:01:01       ^
13:01:01     
13:01:01     AssertionError [ERR_ASSERTION]: 1543428061215n !== 1543428061216n
13:01:01     key=atimeMs, val=1543428061216
13:01:01         at verifyStats (/home/iojs/build/workspace/node-test-commit-smartos/nodes/smartos16-64/test/parallel/test-fs-stat-bigint.js:66:14)
13:01:01         at Object.<anonymous> (/home/iojs/build/workspace/node-test-commit-smartos/nodes/smartos16-64/test/parallel/test-fs-stat-bigint.js:101:3)
13:01:01         at Module._compile (internal/modules/cjs/loader.js:722:30)
13:01:01         at Object.Module._extensions..js (internal/modules/cjs/loader.js:733:10)
13:01:01         at Module.load (internal/modules/cjs/loader.js:620:32)
13:01:01         at tryModuleLoad (internal/modules/cjs/loader.js:560:12)
13:01:01         at Function.Module._load (internal/modules/cjs/loader.js:552:3)
13:01:01         at Function.Module.runMain (internal/modules/cjs/loader.js:775:12)
13:01:01         at startup (internal/bootstrap/node.js:300:19)
13:01:01         at bootstrapNodeJSCore (internal/bootstrap/node.js:826:3)

@joyeecheung
Copy link
Member

Considering some of the tests there are synchronous and some of them are asynchronous, and it fails with python2 tools/test.py --repeat=100 -J parallel/test-fs-stat-bigint, and that the stat calls share one global AliasedBuffer to transport data from C++, it may help to either split this test, or use different AliasedBuffers in the implementation (it's not yet clear to me wether sharing the buffer in callback/synchronous APIs would actually introduce a bug, theoretically it shouldn't)

@bnoordhuis
Copy link
Member

I don't think that's it. There's no sharing taking place, the typed array is passed from C++ to JS and immediately converted to a Stats object.

FWIW, I can reproduce when I run while true; do cat test/.tmp*/* > /dev/null; done in another terminal (but check with mount that the fs isn't mounted with noatime.)

@Trott
Copy link
Member

Trott commented Dec 31, 2018

Another failure on SmartOS, seems off-by-one as in @bnoordhuis's example posted above. Is that a bug in the result or should the test actually permit off-by-one in the timestamp? (Seems like a bug but maybe there's a subtlety I'm missing. I haven't looked closely at all.)

https://ci.nodejs.org/job/node-test-commit-smartos/22861/nodes=smartos16-64/console

test-joyent-smartos16-x64-1

00:12:30 not ok 691 parallel/test-fs-stat-bigint
00:12:30   ---
00:12:30   duration_ms: 0.535
00:12:30   severity: fail
00:12:30   exitcode: 1
00:12:30   stack: |-
00:12:30     (node:636936) ExperimentalWarning: The fs.promises API is experimental
00:12:30     assert.js:86
00:12:30       throw new AssertionError(obj);
00:12:30       ^
00:12:30     
00:12:30     AssertionError [ERR_ASSERTION]: 1546243950493n !== 1546243950494n
00:12:30     key=atimeMs, val=1546243950494
00:12:30         at verifyStats (/home/iojs/build/workspace/node-test-commit-smartos/nodes/smartos16-64/test/parallel/test-fs-stat-bigint.js:66:14)
00:12:30         at fs.stat (/home/iojs/build/workspace/node-test-commit-smartos/nodes/smartos16-64/test/parallel/test-fs-stat-bigint.js:109:7)
00:12:30         at FSReqCallback.oncomplete (fs.js:160:5)
00:12:30   ...

@nodejs/platform-smartos @nodejs/fs

@Trott
Copy link
Member

Trott commented Jan 15, 2019

Failure on debian9-64:

https://ci.nodejs.org/job/node-test-commit-linux/24665/nodes=debian9-64/console

test-softlayer-debian9-x64-1

00:05:02 not ok 593 parallel/test-fs-stat-bigint
00:05:02   ---
00:05:02   duration_ms: 0.141
00:05:02   severity: fail
00:05:02   exitcode: 1
00:05:02   stack: |-
00:05:02     assert.js:86
00:05:02       throw new AssertionError(obj);
00:05:02       ^
00:05:02     
00:05:02     AssertionError [ERR_ASSERTION]: 1547539502126n !== 1547539502127n
00:05:02     key=atimeMs, val=1547539502127
00:05:02         at verifyStats (/home/iojs/build/workspace/node-test-commit-linux/nodes/debian9-64/test/parallel/test-fs-stat-bigint.js:66:14)
00:05:02         at Object.<anonymous> (/home/iojs/build/workspace/node-test-commit-linux/nodes/debian9-64/test/parallel/test-fs-stat-bigint.js:93:3)
00:05:02         at Module._compile (internal/modules/cjs/loader.js:722:30)
00:05:02         at Object.Module._extensions..js (internal/modules/cjs/loader.js:733:10)
00:05:02         at Module.load (internal/modules/cjs/loader.js:621:32)
00:05:02         at tryModuleLoad (internal/modules/cjs/loader.js:564:12)
00:05:02         at Function.Module._load (internal/modules/cjs/loader.js:556:3)
00:05:02         at Function.Module.runMain (internal/modules/cjs/loader.js:775:12)
00:05:02         at executeUserCode (internal/bootstrap/node.js:433:15)
00:05:02         at startExecution (internal/bootstrap/node.js:370:3)
00:05:02   ...

@refack
Copy link
Contributor

refack commented Jan 30, 2019

OS - Windows
Worker - https://ci.nodejs.org/computer/test-azure_msft-win2016-x64-2/
Link - https://ci.nodejs.org/job/node-test-binary-windows/23418/COMPILED_BY=vs2017,RUNNER=win2016,RUN_SUBSET=1/

AssertionError [ERR_ASSERTION]: 1548817600661n !== 1548817600662n
key=atimeMs, val=1548817600662
    at verifyStats (c:\workspace\node-test-binary-windows\test\parallel\test-fs-stat-bigint.js:66:14)
    at c:\workspace\node-test-binary-windows\test\parallel\test-fs-stat-bigint.js:140:3

This is the test code:

(async function() {
const filename = getFilename();
const bigintStats = await promiseFs.stat(filename, { bigint: true });
const numStats = await promiseFs.stat(filename);
verifyStats(bigintStats, numStats);
})();

...
function verifyStats(bigintStats, numStats) {

...
} else if (Number.isSafeInteger(val)) {
assert.strictEqual(
bigintStats[key], BigInt(val),
`${inspect(bigintStats[key])} !== ${inspect(BigInt(val))}\n` +
`key=${key}, val=${val}`
);

So bigintStats is calculated first, before numStats

Is the "atime is changed between the calls" hypothesis is correct, flipping L138 and L139 should flip the sides on the inequality. I'm testing that now.

@refack
Copy link
Contributor

refack commented Jan 30, 2019

AssertionError [ERR_ASSERTION]: 1548858177822n !== 1548858177823n
key=atimeMs, val=1548858177823
    at verifyStats (D:\code\node\test\parallel\test-fs-stat-bigint.js:66:14)
    at D:\code\node\test\parallel\test-fs-stat-bigint.js:140:3
Command: D:\code\node\Release\node.exe D:\code\node\test\parallel\test-fs-stat-bigint.js
[00:21|% 100|+  99|-   1]: Done

numStats is still > bigintStats, even when numStats is calculated first, so this might indicate numerical instability.

@joyeecheung
Copy link
Member

@refack That implies the bug lies somewhere in the numeric conversions?

@refack
Copy link
Contributor

refack commented Jan 30, 2019

@refack That implies the bug lies somewhere in the numeric conversions?

So it seems. This might arise from the way we convert nano-seconds to mili-seconds
for the Number variant we use floating point aritmatic (doubles) so we get a fractional reminder:

const double full_nsec = u_nsec / 1000'000.0;

For BigInt we use integer aritmetic that truncates the reminder:
const auto full_nsec = static_cast<uint64_t>(u_nsec) / 1000'000UL;

Is it possible that when the reminder is higher then some threshold, V8 turns it into the next "SafeInteger"

} else if (Number.isSafeInteger(val)) {
assert.strictEqual(
bigintStats[key], BigInt(val),
`${inspect(bigintStats[key])} !== ${inspect(BigInt(val))}\n` +
`key=${key}, val=${val}`
);

so that 1548858177822.9999d => Number(1548858177823)?

P.S. quick REPL play seems like that is so:

> 1548858177822.9999
1548858177823

@refack refack self-assigned this Mar 4, 2019
@bnoordhuis
Copy link
Member

Floating point rounding error was my hunch too when I first looked at this bug report and it's probably a culprit but perhaps not the only one.

Number.isSafeInteger(1548858177822.9999) returns true but Number.isSafeInteger(1543043785204.1536) (from the original report) does not.

That makes sense because 1548858177822.9999 cannot be exactly represented in a 64 bits double (it gets rounded) but 1543043785204.1536 can.

Rounding also wouldn't explain the ~7 ms gap between 1543043785204.1536 and 1543043785197.

@Trott
Copy link
Member

Trott commented Mar 22, 2019

https://ci.nodejs.org/job/node-test-commit-linux-containered/11608/nodes=ubuntu1604_sharedlibs_shared_x64/testReport/(root)/test/parallel_test_fs_stat_bigint/

test-joyent-ubuntu1604_sharedlibs_container-x64-5

assert.js:87
  throw new AssertionError(obj);
  ^

AssertionError [ERR_ASSERTION]: 1553279804470n !== 1553279804471n
key=atimeMs, val=1553279804471
    at verifyStats (/home/iojs/build/workspace/node-test-commit-linux-containered/test/parallel/test-fs-stat-bigint.js:63:14)
    at Object.<anonymous> (/home/iojs/build/workspace/node-test-commit-linux-containered/test/parallel/test-fs-stat-bigint.js:81:3)
    at Module._compile (internal/modules/cjs/loader.js:814:30)
    at Object.Module._extensions..js (internal/modules/cjs/loader.js:825:10)
    at Module.load (internal/modules/cjs/loader.js:681:32)
    at tryModuleLoad (internal/modules/cjs/loader.js:613:12)
    at Function.Module._load (internal/modules/cjs/loader.js:605:3)
    at Function.Module.runMain (internal/modules/cjs/loader.js:877:12)
    at internal/main/run_main_module.js:21:11

@Trott
Copy link
Member

Trott commented Mar 23, 2019

https://ci.nodejs.org/job/node-test-commit-smartos/24730/nodes=smartos17-64/console

test-joyent-smartos17-x64-2

15:10:28 not ok 708 parallel/test-fs-stat-bigint
15:10:28   ---
15:10:28   duration_ms: 0.425
15:10:28   severity: fail
15:10:28   exitcode: 1
15:10:28   stack: |-
15:10:28     assert.js:87
15:10:28       throw new AssertionError(obj);
15:10:28       ^
15:10:28     
15:10:28     AssertionError [ERR_ASSERTION]: 1553292628257n !== 1553292628258n
15:10:28     key=mtimeMs, val=1553292628258
15:10:28         at verifyStats (/home/iojs/build/workspace/node-test-commit-smartos/nodes/smartos17-64/test/parallel/test-fs-stat-bigint.js:63:14)
15:10:28         at Object.<anonymous> (/home/iojs/build/workspace/node-test-commit-smartos/nodes/smartos17-64/test/parallel/test-fs-stat-bigint.js:98:3)
15:10:28         at Module._compile (internal/modules/cjs/loader.js:814:30)
15:10:28         at Object.Module._extensions..js (internal/modules/cjs/loader.js:825:10)
15:10:28         at Module.load (internal/modules/cjs/loader.js:681:32)
15:10:28         at tryModuleLoad (internal/modules/cjs/loader.js:613:12)
15:10:28         at Function.Module._load (internal/modules/cjs/loader.js:605:3)
15:10:28         at Function.Module.runMain (internal/modules/cjs/loader.js:877:12)
15:10:28         at internal/main/run_main_module.js:21:11
15:10:28   ...

@refack
Copy link
Contributor

refack commented Mar 23, 2019

I have a WIP that refactors the Stats mechanism to simplify it quite a bit, and solve this on the way. Need to find some time to finish it...

@Trott
Copy link
Member

Trott commented Apr 6, 2019

https://ci.nodejs.org/job/node-test-commit-linux-containered/12008/nodes=ubuntu1604_sharedlibs_debug_x64/console

test-softlayer-ubuntu1604_sharedlibs_container-x64-5

00:07:18 not ok 735 parallel/test-fs-stat-bigint
00:07:18   ---
00:07:18   duration_ms: 0.431
00:07:18   severity: fail
00:07:18   exitcode: 1
00:07:18   stack: |-
00:07:18     /home/iojs/build/workspace/node-test-commit-linux-containered/test/common/index.js:681
00:07:18     const crashOnUnhandledRejection = (err) => { throw err; };
00:07:18                                                  ^
00:07:18     
00:07:18     AssertionError [ERR_ASSERTION]: 1554534437999n !== 1554534438000n
00:07:18     key=atimeMs, val=1554534438000
00:07:18         at verifyStats (/home/iojs/build/workspace/node-test-commit-linux-containered/test/parallel/test-fs-stat-bigint.js:63:14)
00:07:18         at /home/iojs/build/workspace/node-test-commit-linux-containered/test/parallel/test-fs-stat-bigint.js:147:5
00:07:18   ...

@Trott
Copy link
Member

Trott commented Jun 10, 2019

No surprise, but still happening, just so no one gets tempted to close this quite yet:

https://ci.nodejs.org/job/node-test-binary-arm-12+/945/RUN_SUBSET=3,label=pi2-docker/console

10:06:58 not ok 178 parallel/test-fs-stat-bigint
10:06:59   ---
10:06:59   duration_ms: 2.958
10:06:59   severity: fail
10:06:59   exitcode: 1
10:06:59   stack: |-
10:06:59     assert.js:89
10:06:59       throw new AssertionError(obj);
10:06:59       ^
10:06:59     
10:06:59     AssertionError [ERR_ASSERTION]: 1560186418390n !== 1560186418391n
10:06:59     key=mtimeMs, val=1560186418391
10:06:59         at verifyStats (/home/iojs/build/workspace/node-test-binary-arm/test/parallel/test-fs-stat-bigint.js:63:14)
10:06:59         at /home/iojs/build/workspace/node-test-binary-arm/test/parallel/test-fs-stat-bigint.js:106:7
10:06:59         at FSReqCallback.oncomplete (fs.js:166:5) {
10:06:59       generatedMessage: false,
10:06:59       code: 'ERR_ASSERTION',
10:06:59       actual: 1560186418390n,
10:06:59       expected: 1560186418391n,
10:06:59       operator: 'strictEqual'
10:06:59     }
10:06:59   ...

Trott added a commit to Trott/io.js that referenced this issue Jun 10, 2019
@joyeecheung
Copy link
Member

I am wondering whether #21387 may help - this moves the timespec-to-ms calculation into JS land entirely. It may make a difference for the precision loss.

Trott added a commit to Trott/io.js that referenced this issue Jun 12, 2019
Refs: nodejs#24593
PR-URL: nodejs#28156
Reviewed-By: Luigi Pinca <[email protected]>
Reviewed-By: Trivikram Kamat <[email protected]>
Reviewed-By: Colin Ihrig <[email protected]>
Reviewed-By: Richard Lau <[email protected]>
@Trott
Copy link
Member

Trott commented Jun 14, 2019

I am wondering whether #21387 may help - this moves the timespec-to-ms calculation into JS land entirely. It may make a difference for the precision loss.

Stress test against master: https://ci.nodejs.org/job/node-stress-single-test/2225/

Stress test against #21387: https://ci.nodejs.org/job/node-stress-single-test/2226/

Might have to do this a few times to figure out how to trigger it in a stress test, but maybe not.

BridgeAR pushed a commit that referenced this issue Jun 17, 2019
Refs: #24593
PR-URL: #28156
Reviewed-By: Luigi Pinca <[email protected]>
Reviewed-By: Trivikram Kamat <[email protected]>
Reviewed-By: Colin Ihrig <[email protected]>
Reviewed-By: Richard Lau <[email protected]>
@Trott
Copy link
Member

Trott commented Feb 7, 2020

This test doesn't seem to be failing anymore. (I used ncu-ci walk to look for failures in CI and also ran the test a few thousand times locally with no failures.) I'm going to close this, but feel free to re-open if this is observed again and/or if a reliable reproduction is found.

@Trott Trott closed this as completed Feb 7, 2020
@Trott
Copy link
Member

Trott commented Feb 8, 2020

Stress test on FreeBSD shows this is still A Thing.

@Trott Trott reopened this Feb 8, 2020
@Trott
Copy link
Member

Trott commented Feb 8, 2020

One failure out of 1000 runs.

https://ci.nodejs.org/view/Stress/job/node-stress-single-test/nodes=freebsd11-x64/41/console

13:41:12 not ok 1 parallel/test-fs-stat-bigint
13:41:12   ---
13:41:12   duration_ms: 0.187
13:41:12   severity: fail
13:41:12   exitcode: 1
13:41:12   stack: |-
13:41:12     assert.js:102
13:41:12       throw new AssertionError(obj);
13:41:12       ^
13:41:12     
13:41:12     AssertionError [ERR_ASSERTION]: 1n !== 9n
13:41:12     key=blocks, val=9
13:41:12         at verifyStats (/usr/home/iojs/build/workspace/node-stress-single-test/nodes/freebsd11-x64/test/parallel/test-fs-stat-bigint.js:83:14)
13:41:12         at /usr/home/iojs/build/workspace/node-stress-single-test/nodes/freebsd11-x64/test/parallel/test-fs-stat-bigint.js:126:7
13:41:12         at FSReqCallback.oncomplete (fs.js:175:5) {
13:41:12       generatedMessage: false,
13:41:12       code: 'ERR_ASSERTION',
13:41:12       actual: 1n,
13:41:12       expected: 9n,
13:41:12       operator: 'strictEqual'
13:41:12     }
13:41:12   ...

Trott pushed a commit to duncanhealy/node that referenced this issue Feb 9, 2020
Change test limit for atime from 2ms to 5ms. Add comment explaining why
the wiggle room is needed.

Fixes: nodejs#24593
@Trott Trott closed this as completed in acfb78b Feb 11, 2020
codebytere pushed a commit that referenced this issue Feb 17, 2020
Change test limit for atime from 2ms to 5ms. Add comment explaining why
the wiggle room is needed.

Fixes: #24593

PR-URL: #30437
Reviewed-By: Anna Henningsen <[email protected]>
Reviewed-By: Denys Otrishko <[email protected]>
Reviewed-By: Rich Trott <[email protected]>
codebytere pushed a commit that referenced this issue Mar 15, 2020
Change test limit for atime from 2ms to 5ms. Add comment explaining why
the wiggle room is needed.

Fixes: #24593

PR-URL: #30437
Reviewed-By: Anna Henningsen <[email protected]>
Reviewed-By: Denys Otrishko <[email protected]>
Reviewed-By: Rich Trott <[email protected]>
codebytere pushed a commit that referenced this issue Mar 17, 2020
Change test limit for atime from 2ms to 5ms. Add comment explaining why
the wiggle room is needed.

Fixes: #24593

PR-URL: #30437
Reviewed-By: Anna Henningsen <[email protected]>
Reviewed-By: Denys Otrishko <[email protected]>
Reviewed-By: Rich Trott <[email protected]>
codebytere pushed a commit that referenced this issue Mar 30, 2020
Change test limit for atime from 2ms to 5ms. Add comment explaining why
the wiggle room is needed.

Fixes: #24593

PR-URL: #30437
Reviewed-By: Anna Henningsen <[email protected]>
Reviewed-By: Denys Otrishko <[email protected]>
Reviewed-By: Rich Trott <[email protected]>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
flaky-test Issues and PRs related to the tests with unstable failures on the CI. fs Issues and PRs related to the fs subsystem / file system. macos Issues and PRs related to the macOS platform / OSX.
Projects
None yet
Development

Successfully merging a pull request may close this issue.

6 participants