Lincheck tests are slow when the number of CPUs is lower than the number of parallel threads

Question

Lincheck tests are slow when the number of CPUs is lower than the number of parallel threads

ndkoval opened this issue 7 months ago · 5 comments

Please see #278 (comment) for details:

It seems to have hung on aarch64.

I have run it successfully locally on an M3 macbook, but not on CircleCI, likely because of fewer cpu cores (as we saw that affected Lincheck in past bugs).

Answer 1 · 2024-02-22T19:23:36.000Z

Ok, I think now I am able to reproduce this issue locally using a docker container with a restricted number of CPUs.
I am investigating the root cause of the issue now.

Answer 2 · 2024-03-16T18:25:56.000Z

Thanks, working beautifully now!

Answer 3 · 2024-03-23T02:21:08.000Z

was there a regression? I tried to upgrade 2.27 to 2.28.x and it times out (2.28, 2.28.1) like before.

Answer 4 · 2024-03-25T11:49:47.000Z

Hi @ben-manes !
Yes, this is probably a regression. I'll take a look at it.
Could you please pin the version 2.27 for now?

Answer 5 · 2024-03-25T15:20:31.000Z

thanks @eupp! Please consider adding a test case.