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

Investigate parallel/test-cluster-send-handle-large-payload failure on FreeBSD #14844

Closed
Trott opened this issue Aug 15, 2017 · 10 comments
Closed
Labels
cluster Issues and PRs related to the cluster subsystem. flaky-test Issues and PRs related to the tests with unstable failures on the CI. freebsd Issues and PRs related to the FreeBSD platform.

Comments

@Trott
Copy link
Member

Trott commented Aug 15, 2017

  • Version: 9.0.0-pre
  • Platform: freebsd10-64
  • Subsystem: test

New and different failure for test-cluster-send-handle-large-payload, this time on FreeBSD.

https://ci.nodejs.org/job/node-test-commit-freebsd/10969/nodes=freebsd10-64/console

not ok 243 parallel/test-cluster-send-handle-large-payload
  ---
  duration_ms: 1.638
  severity: fail
  stack: |-
    undefined:1
    aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
    
    SyntaxError: Unexpected token a in JSON at position 0
        at JSON.parse (<anonymous>)
        at Pipe.channel.onread (internal/child_process.js:479:28)
  ...

@addaleax

@Trott Trott added flaky-test Issues and PRs related to the tests with unstable failures on the CI. cluster Issues and PRs related to the cluster subsystem. freebsd Issues and PRs related to the FreeBSD platform. labels Aug 15, 2017
@addaleax addaleax self-assigned this Aug 15, 2017
@addaleax
Copy link
Member

sigh

Not sure how this can happen, but at least this is an OS I can get access to.

@addaleax
Copy link
Member

Stress test to get an idea of just how flaky this is: https://ci.nodejs.org/job/node-stress-single-test/1375/nodes=freebsd10-64/

@BridgeAR
Copy link
Member

It happened again recently.

@addaleax addaleax removed their assignment Dec 8, 2017
@MylesBorins
Copy link
Contributor

@Trott
Copy link
Member Author

Trott commented Mar 12, 2018

@Trott
Copy link
Member Author

Trott commented Mar 12, 2018

Stress test, once in sequence and once with -J for parallelism.

Sequence: https://ci.nodejs.org/job/node-stress-single-test/1787/nodes=freebsd11-x64/
Parallel: https://ci.nodejs.org/job/node-stress-single-test/1788/nodes=freebsd11-x64/

@Trott
Copy link
Member Author

Trott commented Mar 12, 2018

I messed up the parallel stress test. Here it is again, hopefully correctly done this time:

https://ci.nodejs.org/job/node-stress-single-test/1789/nodes=freebsd11-x64/

@Trott
Copy link
Member Author

Trott commented Mar 13, 2018

Parallel results suggest this is an issue about resource constraints and it might be solved (or mitigated) by moving the test to sequential. I'll submit a PR to do that.

ok 1 parallel/test-cluster-send-handle-large-payload
  ---
  duration_ms: 2.853
  ...
ok 2 parallel/test-cluster-send-handle-large-payload
  ---
  duration_ms: 3.294
  ...
ok 3 parallel/test-cluster-send-handle-large-payload
  ---
  duration_ms: 3.316
  ...
ok 4 parallel/test-cluster-send-handle-large-payload
  ---
  duration_ms: 3.172
  ...
not ok 5 parallel/test-cluster-send-handle-large-payload
  ---
  duration_ms: 3.92
  severity: fail
  stack: |-
    undefined:1
    aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
    
    SyntaxError: Unexpected token a in JSON at position 0
        at JSON.parse (<anonymous>)
        at Pipe.channel.onread (internal/child_process.js:490:28)
  ...
ok 6 parallel/test-cluster-send-handle-large-payload
  ---
  duration_ms: 3.236

Trott added a commit to Trott/io.js that referenced this issue Mar 13, 2018
test-cluster-send-handle-large-payload is susceptible to failure if it
is competing for resources. Move to `sequential` directory so it is not
competing with other tests.

Fixes: nodejs#14844
@Trott
Copy link
Member Author

Trott commented Mar 13, 2018

(Sequential stress test turned up green. It was just the parallel stress test that showed the failures.)

@santigimeno
Copy link
Member

I looked into this a while ago and IIRC it seemed to me that it could be related with a bug in the freeBSD kernel where data was silently lost when sending SCM_RIGHTS messages with large payload. See: https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=181741 and https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=215933.

@Trott Trott closed this as completed in 5e23b65 Mar 15, 2018
targos pushed a commit that referenced this issue Mar 17, 2018
test-cluster-send-handle-large-payload is susceptible to failure if it
is competing for resources. Move to `sequential` directory so it is not
competing with other tests.

Fixes: #14844

PR-URL: #19311
Reviewed-By: Daniel Bevenius <daniel.bevenius@gmail.com>
MylesBorins pushed a commit that referenced this issue Mar 20, 2018
test-cluster-send-handle-large-payload is susceptible to failure if it
is competing for resources. Move to `sequential` directory so it is not
competing with other tests.

Fixes: #14844

PR-URL: #19311
Reviewed-By: Daniel Bevenius <daniel.bevenius@gmail.com>
MylesBorins pushed a commit that referenced this issue Jul 31, 2018
test-cluster-send-handle-large-payload is susceptible to failure if it
is competing for resources. Move to `sequential` directory so it is not
competing with other tests.

Fixes: #14844

PR-URL: #19311
Reviewed-By: Daniel Bevenius <daniel.bevenius@gmail.com>
rvagg pushed a commit that referenced this issue Aug 16, 2018
test-cluster-send-handle-large-payload is susceptible to failure if it
is competing for resources. Move to `sequential` directory so it is not
competing with other tests.

Fixes: #14844

PR-URL: #19311
Reviewed-By: Daniel Bevenius <daniel.bevenius@gmail.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
cluster Issues and PRs related to the cluster subsystem. flaky-test Issues and PRs related to the tests with unstable failures on the CI. freebsd Issues and PRs related to the FreeBSD platform.
Projects
None yet
Development

Successfully merging a pull request may close this issue.

5 participants