kubernetes-csi-livenessprobe

A sidecar container that can be included in a CSI plugin pod to enable integration with Kubernetes Liveness Probe.

Download this Image

The image is available on cgr.dev:

docker pull cgr.dev/chainguard/kubernetes-csi-livenessprobe:latest

Run it

Generally speaking, the kubernetes-csi-livenessprobe is a low level Kubernetes component not meant to be managed directly. However, all the steps outlined in the upstream repo apply just as well to the Chainguard Image version.

Test it

A simple smoke test is provided below that leverages docker to spin up the liveness probe attached to a csi driver and ensure basic functionality is met:

#!/usr/bin/env bash

set -o errexit -o nounset -o errtrace -o pipefail -x

if [[ "${IMAGE_NAME}" == "" ]]; then
	echo "Must set IMAGE_NAME environment variable. Exiting."
	exit 1
fi

UDS="/csi/csi.sock"
CSI_ENDPOINT="unix:/$UDS"

# Start hostpathplugin in the background
docker run -d --name csi-driver -v $(pwd)/csi:/csi quay.io/k8scsi/hostpathplugin:v1.6.0 --endpoint=$CSI_ENDPOINT -nodeid 1 --v=5
trap "docker rm -f csi-driver" EXIT

docker run -d --name probe -p 9808:9808 -v $(pwd)/csi:/csi "$IMAGE_NAME" --v=5 --csi-address=$UDS
trap "docker rm -f probe" EXIT

# Give time to CSI hostpathplugin and livenessprobe to initialize
sleep 3

# Requesting health
health=$(curl -I http://localhost:9808/healthz | grep HTTP | awk '{print $2}')
if [[ "x$health" != "x200" ]]; then
	echo "Health check failed, but it was not supposed to, exiting..."
	exit 1
fi

# Killing hostpathplugin
docker stop csi-driver
sleep 3

# Requesting health, should fail since hostpathplugin is gone
health=$(curl -I http://localhost:9808/healthz | grep HTTP | awk '{print $2}')
if [[ "x$health" != "x500" ]]; then
	echo "Health check did not detect driver failure, returned code: $health, exiting..."
	exit 1
fi

exit 0