systemd cgroup driver detection does not work on servers

[brandond]

It looks like this was regressed by a9b5a1933fb. On servers, the NOTIFY_SOCKET environment variable gets unset, which prevents the cgroup detection code from detecting that it is running under systemd.

Originally posted by @brandond in #5454 (comment)

[est-suse]

Environment Details

Infrastructure

• Cloud
• Hosted

Node(s) CPU architecture, OS, and Version:

Linux ip-172-31-35-90 5.10.0-14-cloud-amd64 #1 SMP Debian 5.10.113-1 (2022-04-29) x86_64 GNU/Linux
PRETTY_NAME="Debian GNU/Linux 11 (bullseye)"
NAME="Debian GNU/Linux"
VERSION_ID="11"
VERSION="11 (bullseye)"
VERSION_CODENAME=bullseye
ID=debian
HOME_URL="https://www.debian.org/"
SUPPORT_URL="https://www.debian.org/support"
BUG_REPORT_URL="https://bugs.debian.org/"

Cluster Configuration:
1 server

Config.yaml:

N/A

Testing Steps

1. Copy config.yaml

install K3s: curl -sfL https://get.k3s.io | sh -s - --write-kubeconfig-mode 644 --disable traefik --kubelet-arg cgroup-driver=systemd

The pods will fail after 4-5 minutes:

Validation Results:

• k3s version used for validation:

Validated with version: v1.24.4-rc1+k3s1

Install K3s with the command below:

curl -sfL https://get.k3s.io | INSTALL_K3S_VERSION=v1.24.4-rc1+k3s1 sh -s - --write-kubeconfig-mode 644 --disable traefik --kubelet-arg cgroup-driver=systemd

After checking for 60 min, Pods never failed.