Uh oh!
There was an error while loading. Please reload this page.
test: deflake test-http-regr-gh-2928 - #49574
Conversation
This comment was marked as outdated.
This comment was marked as outdated.
nodejs-github-bot
commented
Sep 9, 2023
nodejs-github-bot
commented
Sep 9, 2023
lpinca
commented
Sep 9, 2023
It is still flaky: https://ci.nodejs.org/view/Stress/job/node-stress-single-test/426/. |
lpinca
commented
Sep 10, 2023
It seems that data stops flowing in the FreeBSD 12 machine where the stress test is run. Here is an example https://ci.nodejs.org/view/Stress/job/node-stress-single-test/429/. I don't know why. I can't reproduce the issue in a locally installed FreeBSD 12 VM. |
f06e780 to
68fcb02Compare6250590 to
a2f0ca3CompareHard code the value of the host parameter to `common.localhostIPv4` in `server.listen()` and `net.connect()`. This 1. ensures that the client `socket._handle` is not reinitialized during connection due to the family autodetection algorithm, preventing `parser.consume()` from being called with an invalid `socket._handle` parameter. 2. works around an issue in the FreeBSD 12 machine where the stress test is run where some sockets get stuck after connection. Closes: nodejs#49565Fixes: nodejs#49564
a2f0ca3 to
6db17f7Comparelpinca
commented
Sep 10, 2023
I now occasionally get a on macOS but that can probably be fixed by reducing the number of sockets. The test does not need to create 1k sockets. Let's see how it goes. |
The reason why some sockets get stuck with no data flowing on the FreeBSD 12 machine when the host is not specified should be investigated (faulty IPv6?, buggy family autodetection?) but I think it should not block this. |
lpinca
commented
Sep 10, 2023
FreeBSD 12 stress test: https://ci.nodejs.org/view/Stress/job/node-stress-single-test/436/ |
nodejs-github-bot
commented
Sep 10, 2023
nodejs-github-bot
commented
Sep 10, 2023
nodejs-github-bot
commented
Sep 10, 2023
lpinca
commented
Sep 15, 2023
cc: @nodejs/testing |
joyeecheung
commented
Sep 16, 2023
Is it in shape for review now? (Can’t say for sure from the thread) |
lpinca
commented
Sep 16, 2023
@joyeecheung yes. |
nodejs-github-bot
commented
Sep 18, 2023
Commit Queue failed- Loading data for nodejs/node/pull/49574 ✔ Done loading data for nodejs/node/pull/49574 ----------------------------------- PR info ------------------------------------ Title test: deflake test-http-regr-gh-2928 (#49574) ⚠ Could not retrieve the email or name of the PR author's from user's GitHub profile! Branch lpinca:deflake/test-http-regr-gh-2928 -> nodejs:main Labels test, needs-ci Commits 1 - test: deflake test-http-regr-gh-2928 Committers 1 - Luigi Pinca PR-URL: https://github.com/nodejs/node/pull/49574 Fixes: https://github.com/nodejs/node/pull/49565 Fixes: https://github.com/nodejs/node/issues/49564 Reviewed-By: Yagiz Nizipli Reviewed-By: James M Snell ------------------------------ Generated metadata ------------------------------ PR-URL: https://github.com/nodejs/node/pull/49574 Fixes: https://github.com/nodejs/node/pull/49565 Fixes: https://github.com/nodejs/node/issues/49564 Reviewed-By: Yagiz Nizipli Reviewed-By: James M Snell -------------------------------------------------------------------------------- ℹ This PR was created on Sat, 09 Sep 2023 12:59:22 GMT ✔ Approvals: 2 ✔ - Yagiz Nizipli (@anonrig) (TSC): https://github.com/nodejs/node/pull/49574#pullrequestreview-1618710274 ✔ - James M Snell (@jasnell) (TSC): https://github.com/nodejs/node/pull/49574#pullrequestreview-1629983759 ✔ Last GitHub CI successful ℹ Last Full PR CI on 2023-09-10T19:40:28Z: https://ci.nodejs.org/job/node-test-pull-request/53872/ - Querying data for job/node-test-pull-request/53872/ ✔ Last Jenkins CI successful -------------------------------------------------------------------------------- ✔ No git cherry-pick in progress ✔ No git am in progress ✔ No git rebase in progress -------------------------------------------------------------------------------- - Bringing origin/main up to date... From https://github.com/nodejs/node * branch main -> FETCH_HEAD 5c39ee6f87..b651e37d2e main -> origin/main ✔ origin/main is now up-to-date main is out of sync with origin/main. Mismatched commits: - b651e37d2e 2023-09-18, Version 20.7.0 (Current) -------------------------------------------------------------------------------- HEAD is now at b651e37d2e 2023-09-18, Version 20.7.0 (Current) ✔ Reset to origin/main - Downloading patch for 49574 From https://github.com/nodejs/node * branch refs/pull/49574/merge -> FETCH_HEAD ✔ Fetched commits as b651e37d2e36..6db17f7c5518 -------------------------------------------------------------------------------- [main 26c4258a63] test: deflake test-http-regr-gh-2928 Author: Luigi Pinca Date: Sat Sep 9 14:44:31 2023 +0200 1 file changed, 2 insertions(+), 2 deletions(-) ✔ Patches applied -------------------------------------------------------------------------------- ⚠ Found Fixes: https://github.com/nodejs/node/issues/49564, skipping.. --------------------------------- New Message ---------------------------------- test: deflake test-http-regr-gh-2928https://github.com/nodejs/node/actions/runs/6226235586 |
Hard code the value of the host parameter to `common.localhostIPv4` in `server.listen()` and `net.connect()`. This 1. ensures that the client `socket._handle` is not reinitialized during connection due to the family autodetection algorithm, preventing `parser.consume()` from being called with an invalid `socket._handle` parameter. 2. works around an issue in the FreeBSD 12 machine where the stress test is run where some sockets get stuck after connection. PR-URL: #49574Closes: #49565Fixes: #49564 Reviewed-By: Yagiz Nizipli <yagiz@nizipli.com> Reviewed-By: James M Snell <jasnell@gmail.com>
lpinca
commented
Sep 18, 2023
Landed in 18e00a5. |
Hard code the value of the host parameter to `common.localhostIPv4` in `server.listen()` and `net.connect()`. This 1. ensures that the client `socket._handle` is not reinitialized during connection due to the family autodetection algorithm, preventing `parser.consume()` from being called with an invalid `socket._handle` parameter. 2. works around an issue in the FreeBSD 12 machine where the stress test is run where some sockets get stuck after connection. PR-URL: #49574Closes: #49565Fixes: #49564 Reviewed-By: Yagiz Nizipli <yagiz@nizipli.com> Reviewed-By: James M Snell <jasnell@gmail.com>
Hard code the value of the host parameter to `common.localhostIPv4` in `server.listen()` and `net.connect()`. This 1. ensures that the client `socket._handle` is not reinitialized during connection due to the family autodetection algorithm, preventing `parser.consume()` from being called with an invalid `socket._handle` parameter. 2. works around an issue in the FreeBSD 12 machine where the stress test is run where some sockets get stuck after connection. PR-URL: #49574 Backport-PR-URL: #52384Closes: #49565Fixes: #49564 Reviewed-By: Yagiz Nizipli <yagiz@nizipli.com> Reviewed-By: James M Snell <jasnell@gmail.com>
Hard code the value of the host parameter to `common.localhostIPv4` in `server.listen()` and `net.connect()`. This 1. ensures that the client `socket._handle` is not reinitialized during connection due to the family autodetection algorithm, preventing `parser.consume()` from being called with an invalid `socket._handle` parameter. 2. works around an issue in the FreeBSD 12 machine where the stress test is run where some sockets get stuck after connection. PR-URL: nodejs/node#49574Closes: nodejs/node#49565Fixes: nodejs/node#49564 Reviewed-By: Yagiz Nizipli <yagiz@nizipli.com> Reviewed-By: James M Snell <jasnell@gmail.com>
Hard code the value of the host parameter to
common.localhostIPv4inserver.listen()andnet.connect(). Thissocket._handleis not reinitialized duringconnection due to the family autodetection algorithm, preventing
parser.consume()from being called with an invalidsocket._handleparameter.
is run where some sockets get stuck after connection.
Closes: #49565
Fixes: #49564