# Ingress controller reported: epoll_create () failed (24: Too many open files)

Ingress controller does not work with a POWER machine, which has 160 cores. Ingress controller might fail when it is running on a node with lots of cores.

Causes
Ingress controller might be running on a node that has too many cores. The maximum number of open file descriptors is calculated with the following formula: *RLIMIT_NOFILE/worker-processes) - 1024. To resolve, you can either decrease the value of the worker processes, or increase the value of the RLIMIT_NOFILE of the container.

Solution one: Edit the configMap of nginx-ingress-controller with a decreased value of worker-processes.
To edit the configmap of nginx-ingress-controller, run the following command:

kubectl -n kube-system edit cm nginx-ingress-controller

Add worker-processes: "2" to the configMap, as it is in the following example. Note: The value might not be 2, depending on your sysctl configuration.

# Edit the following object. Lines beginning with a '#' are ignored,
# and an empty file will abort the edit. If an error occurs while saving this file will be
# reopened with the relevant failures.
#
apiVersion: v1
data:
body-size: "0"
disable-access-log: "true"
worker-processes: "2"

What happened?
Just deploy a fresh k8s cluster with crio, deploy ingress-nginx and try to load ingresses. Nginx throws errors because it cannot respawn its worker processes:

2022/04/05 09:27:49 [alert] 56#56: sendmsg() failed (9: Bad file descriptor)
2022/04/05 09:27:50 [alert] 56#56: worker process 1191 exited with fatal code 2 and cannot be respawned