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

feature: configuration of pids_limit (podman) #19820

Open
rriemann opened this issue Oct 20, 2024 · 3 comments
Open

feature: configuration of pids_limit (podman) #19820

rriemann opened this issue Oct 20, 2024 · 3 comments
Labels
kind/support Categorizes issue or PR as a support question. triage/needs-information Indicates an issue needs more information in order to work on it.

Comments

@rriemann
Copy link

rriemann commented Oct 20, 2024

What Happened?

I use minikube in two ways:

  • minikube start --driver=podman --container-runtime=cri-o --cpus="4" --memory="4096"
  • minikube start --driver=podman --container-runtime=containerd --cpus="4" --memory="4096"

The workloads fail to create threads with error pthread_create failed (EAGAIN). I can fix this issue outside of minikube with:

podman run --pids-limit -1 alpine cat /sys/fs/cgroup/pids.max

However, I am missing a similar option for minikube to set --pids-limit.

Workaround to change the value system-wide

Have this in your /etc/containers/containers.conf:

[containers]
pids_limit=0

source: kubernetes-sigs/kind#2896 (comment)

Attach the log file

I could not find anything on this in the minikube logs. However, dmesg has:

[ 5351.722836] cgroup: fork rejected by pids controller in /libpod_parent/libpod-6136e1615cd65448ba547e2db748f956cb90c27a7fd54665df8b320302debc07/kubepods.slice/kubepods-pode8d39171_af9d_4c07_b03e_6e72aefc26a8.slice/crio-27f14c6154e2630984fad3eb585e75456d10be978b80cb2c587efc77aa80f219.scope

Operating System

Redhat/Fedora

Driver

Podman

@medyagh
Copy link
Member

medyagh commented Oct 21, 2024

@rriemann thanks for creating the issue, would you say it would be a good default to change the PID limit for everyone? or does this issue only affect your cluster with a specific workload ? do you mind sharing more about what your workload is ?

@medyagh
Copy link
Member

medyagh commented Oct 21, 2024

/triage needs-information
/kind support

@k8s-ci-robot k8s-ci-robot added triage/needs-information Indicates an issue needs more information in order to work on it. kind/support Categorizes issue or PR as a support question. labels Oct 21, 2024
@rriemann
Copy link
Author

I am no expert in this area. I was playing with https://github.com/bitnami/charts/tree/main/bitnami/ejbca and just hit the limit on every deploy.

I assume that my OS Fedora limited pids for a reason and if I do not play with minikube, I'd like to conserve the default configuration of my OS.

I understand most people use minikube for testing. So in that sense I would recommend to inrease the default pid limit for everyone. I do not know which value would be a good fit. Hence, I believe the pid limit should be accessible via the arguments to the minikube start command and rest inside the user config next to the existing values for cpus and memory.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
kind/support Categorizes issue or PR as a support question. triage/needs-information Indicates an issue needs more information in order to work on it.
Projects
None yet
Development

No branches or pull requests

3 participants