test: set test-http-regr-gh-2928 as flaky - #50228
Conversation
Sorry, something went wrong.
Sorry, something went wrong.
Sorry, something went wrong.
|
It is flaky only on Windows 2016, I wonder if there is something wrong on the test machine. |
Sorry, something went wrong.
StefanStojanovic
left a comment
There was a problem hiding this comment.
It is flaky only on Windows 2016, I wonder if there is something wrong on the test machine.
I had my suspicions about the x86 build that was tested there, but couldn't reproduce locally. Regarding the machines, those are planned for removal soon, so we can check after that happens to see if it was only machines or something else. Until that happens, I'm OK with marking the test as flaky.
Sorry, something went wrong.
|
@richardlau is it possible to add the the Windows 2016 to the node-stress-single-test job? I want to try this patch diff --git a/test/sequential/test-http-regr-gh-2928.js b/test/sequential/test-http-regr-gh-2928.js
index 25476e0453..f6a9e16032 100644
--- a/test/sequential/test-http-regr-gh-2928.js
+++ b/test/sequential/test-http-regr-gh-2928.js
@@ -8,6 +8,8 @@ const httpCommon = require('_http_common');
const { HTTPParser } = require('_http_common');
const net = require('net');
+httpCommon.parsers.max = 50;
+
const COUNT = httpCommon.parsers.max + 1;
const parsers = new Array(COUNT);
We might want to apply it regardless of whether it solves the issue or not. There is no reason to use 1000 parsers in the test. |
Sorry, something went wrong.
The job already has win2016-vs2017 but since it has Visual Studio 2017 the job won't even schedule due to our versioning rules. In the node-test-commit-windows-fanned job Node.js is compiled for Windows on one machine and then the binary is copied over to other machines (e.g. the win2016 machines) for testing. node-stress-single-test is set up to compile and run on the same machine. AFAICT all of the win2016 machines have vs2017 only. |
Sorry, something went wrong.
|
I understand, thank you. |
Sorry, something went wrong.
|
Fast-track has been requested by @anonrig. Please 👍 to approve. |
Sorry, something went wrong.
|
Can we try with #50240 before landing this? |
Sorry, something went wrong.
The maximum number of parsers in the free list is set to 1000. However the test does not need to use this maximum. Reduce it to 50. Refs: nodejs#50228 (comment)
The maximum number of parsers in the free list is set to 1000. However the test does not need to use this maximum. Reduce it to 50. Refs: #50228 (comment) PR-URL: #50240 Fixes: #49564 Reviewed-By: Yagiz Nizipli <yagiz@nizipli.com> Reviewed-By: Vinícius Lourenço Claro Cardoso <contact@viniciusl.com.br>
The maximum number of parsers in the free list is set to 1000. However the test does not need to use this maximum. Reduce it to 50. Refs: #50228 (comment) PR-URL: #50240 Fixes: #49564 Reviewed-By: Yagiz Nizipli <yagiz@nizipli.com> Reviewed-By: Vinícius Lourenço Claro Cardoso <contact@viniciusl.com.br>
The maximum number of parsers in the free list is set to 1000. However the test does not need to use this maximum. Reduce it to 50. Refs: #50228 (comment) PR-URL: #50240 Fixes: #49564 Reviewed-By: Yagiz Nizipli <yagiz@nizipli.com> Reviewed-By: Vinícius Lourenço Claro Cardoso <contact@viniciusl.com.br>
The maximum number of parsers in the free list is set to 1000. However the test does not need to use this maximum. Reduce it to 50. Refs: #50228 (comment) PR-URL: #50240 Backport-PR-URL: #52384 Fixes: #49564 Reviewed-By: Yagiz Nizipli <yagiz@nizipli.com> Reviewed-By: Vinícius Lourenço Claro Cardoso <contact@viniciusl.com.br>
The maximum number of parsers in the free list is set to 1000. However the test does not need to use this maximum. Reduce it to 50. Refs: nodejs/node#50228 (comment) PR-URL: nodejs/node#50240 Fixes: nodejs/node#49564 Reviewed-By: Yagiz Nizipli <yagiz@nizipli.com> Reviewed-By: Vinícius Lourenço Claro Cardoso <contact@viniciusl.com.br>
Ref: #49564