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

Add an option to automatically set worker_connections based on worker_rlimit_nofile #3604

Merged
merged 1 commit into from Dec 31, 2018

Conversation

@ramnes
Copy link
Contributor

ramnes commented Dec 27, 2018

What this PR does / why we need it:

This PR makes the max-worker-connections option use maxOpenFile by default rather than 16384, and adds a max-worker-open-files option for those who'd want to manually set both.

See #3557 for rationale.

Which issue this PR fixes:

It closes #3557.

Special notes for your reviewer:

First of all, thanks for passing by! Please note that I'm not a Gopher; any remark is more than welcome.

@k8s-ci-robot k8s-ci-robot added the size/M label Dec 27, 2018
@k8s-ci-robot

This comment has been minimized.

Copy link
Contributor

k8s-ci-robot commented Dec 27, 2018

Thanks for your pull request. Before we can look at your pull request, you'll need to sign a Contributor License Agreement (CLA).

📝 Please follow instructions at https://git.k8s.io/community/CLA.md#the-contributor-license-agreement to sign the CLA.

It may take a couple minutes for the CLA signature to be fully registered; after that, please reply here with a new comment and we'll verify. Thanks.


Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository. I understand the commands that are listed here.

@ramnes

This comment has been minimized.

Copy link
Contributor Author

ramnes commented Dec 27, 2018

@aledbf

This comment has been minimized.

Copy link
Member

aledbf commented Dec 31, 2018

/lgtm
/approve

@aledbf

This comment has been minimized.

Copy link
Member

aledbf commented Dec 31, 2018

@ramnes thanks!

@k8s-ci-robot

This comment has been minimized.

Copy link
Contributor

k8s-ci-robot commented Dec 31, 2018

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: aledbf, ramnes

The full list of commands accepted by this bot can be found here.

The pull request process is described here

Needs approval from an approver in each of these files:

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment

@k8s-ci-robot k8s-ci-robot merged commit c3eeaca into kubernetes:master Dec 31, 2018
3 checks passed
3 checks passed
cla/linuxfoundation ramnes authorized
Details
continuous-integration/travis-ci/pr The Travis CI build passed
Details
tide In merge pool.
Details
@aledbf aledbf added this to done in 0.22.0 Dec 31, 2018
@aledbf

This comment has been minimized.

Copy link
Member

aledbf commented Jan 12, 2019

@ramnes I am going to rever the default value for MaxWorkerConnections. Using 0 as default increases the default memory usage ~3.5x something that is not acceptable.

Screenshot of the controller in an empty cluster without ingresses

screenshot from 2019-01-12 10 21 58

@ramnes

This comment has been minimized.

Copy link
Contributor Author

ramnes commented Jan 12, 2019

@aledbf This is really weird, I don't see any reason that it would change the memory usage when idle. Are you sure that it's related to that commit? Also, in your example, what is the value that got generated?

@aledbf

This comment has been minimized.

Copy link
Member

aledbf commented Jan 12, 2019

Are you sure that it's related to that commit?

Yes. The value was 261120 (same than worker_rlimit_nofile). Once I returned the value to the previous level the memory usage returned to the "normal" value

screenshot from 2019-01-12 12 21 26

@ramnes

This comment has been minimized.

Copy link
Contributor Author

ramnes commented Jan 12, 2019

I just tried it myself and I can confirm that Nginx RAM usage on idle and worker_connections are highly correlated. I wouldn't have expected this and it's good to know, thanks for the find! Looks like Nginx is pre-allocating its events space, most likely to avoid the penalty of doing it when scaling up.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
0.22.0
  
done
3 participants
You can’t perform that action at this time.