[CI] Check duration of controller_managers_srv test #1267
Assignees
Comments
|
Hello! I just looked into this issue. It seems like, this is happening due to the the limit of the CPU cores. On my PC, the test runs for 33 seconds However, on the CI it seems to take longer due to the limit in the CPUs and also as they are not so powerful enough. It took more than 120 seconds for around 160+ controllers (more or less 40 controllers remaining to configure). As this is the case, I recommend reducing the number of controllers for the sake for determinism and less run time on Ci |
|
I've also tried to reduce the number of CPUs used by my docker container recommended by as @moriarty. I also see that the runtime of the tests is nearly double. |
|
Can't we improve the speed of the service itself? Or which part is taking such a long time per controller? |
|
It seems to be fixed with #1384 on rolling, There wasn't any failing test due to a timeout in the last weeks on humble/iron, so I hope this is not a problem any more. |
|
@saikishor I analyzed the times for the failing CI in #1700 with #1781 Times are in ms, iron behaves similar to rolling.
#1384 hasn't been backported to iron, why is iron fine but humble not?
|
|
@christophfroehlich if the Iron doesn't have #1384 it doesn't make a lot of sense as you said. Moreover, it's like 6 times. Maybe we can reduce the number for now and see. Moreover, in that test, we have like 100+ controllers, this is not so usual case, so it shouldn't be a lot of problem. What do you say? |
|
I have no idea actually :D You are more familiar with this test, could you have a quick look and compare the test from humble/iron? |
|
Sure, I will check it once I get some free time :) |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
If the test makes sense this shouldn't be a problem. I'm just curious why it takes so long actually? There is no sleep or timing in there? Does the service call take so much time?
Originally posted by @christophfroehlich in #1260 (comment)
The text was updated successfully, but these errors were encountered: