Setup dapr on AWS and calico

May 16, 2022 - 2 minutes read - 286 words

Low resource usage

Our system had low resource usage, however my several pods got stucked in ContainerCreating state. Kubernetes events show that "add cmd: failed to assign an IP address to container". "kubectl top nodes" still shows the usage of resources still low.

jackl@LAPTOP-IIHUF2DR:~$ k --context jack.l@test-dev.ap-southeast-1.eksctl.io top nodes
NAME                                                 CPU(cores)   CPU%   MEMORY(bytes)   MEMORY%
ip-192-168-102-109.ap-southeast-1.compute.internal   143m         7%     3217Mi          45%
ip-192-168-110-171.ap-southeast-1.compute.internal   117m         6%     2892Mi          41%

Migrate to calico

Recreate cluster without nodegroups
delete aws-node daemonsets
apply calico-vxlan
create nodegroups

# v3.20.2
# https://projectcalico.docs.tigera.io/manifests/calico-vxlan.yaml

export AWS_PROFILE=test-dev
eksctl create cluster -f test-dev-cluster.yaml --without-nodegroup
kubectl delete daemonset -n kube-system aws-node
kubectl apply -f calico-vxlan.yaml
eksctl create nodegroup -f test-dev-cluster.yaml

Issue: I don’t see the Dapr sidecar injected to my pod

# https://github.com/dapr/dapr/pull/4179#issue-1119608603
dapr_sidecar_injector:
  hostNetwork: true
  # port conflict etc, need to change port
  healthzPort: 8091

helm repo add dapr https://dapr.github.io/helm-charts/
helm repo update
# See which chart versions are available
helm search repo dapr --devel --versions


#--version is needed
helm upgrade --install dapr dapr/dapr \
  --version=1.7.2 \
  --namespace dapr-system \
  --create-namespace \
  --values values.yaml

# prefer to
dapr init -k \
   --set dapr_sidecar_injector.hostNetwork=true,dapr_sidecar_injector.healthzPort=8091

Debug

//.vscode/launch.json
{
    "configurations": [
        {
            "name": "Launch divideapp with Kubernetes",
            "program": "${workspaceFolder}/app.js",
            "request": "launch",
            "skipFiles": [
                "<node_internals>/**"
            ],
            "type": "node",
            "preLaunchTask": "bridge-to-kubernetes.resource",
            "env": {
                "GRPC_DNS_RESOLVER": "native"
            },


        },
        {
            "name": "Launch divideapp",
            "program": "${workspaceFolder}/app.js",
            "request": "launch",
            "skipFiles": [
                "<node_internals>/**"
            ],
            "type": "node"
        }
    ]
}

//.vscode/tasks.json
{
	"version": "2.0.0",
	"tasks": [
		{
			"label": "bridge-to-kubernetes.resource",
			"type": "bridge-to-kubernetes.resource",
			"resource": "divideapp-dapr",
			"resourceType": "service",
			"ports": [
				4000 //match the port number in code
			],
			"targetCluster": "arn:aws:eks:ap-southeast-1:372827928396:cluster/cnpsp",
			"targetNamespace": "default",
			"useKubernetesServiceEnvironmentVariables": true
		}
	]
}

git clone https://github.com/dapr/quickstarts.git
cd quickstarts/tutorials/distributed-calculator/deploy
kubectl apply -f .

cd node
yarn install
code .

vscode

Debug only when the pod is ready

Low resource usage

Migrate to calico

Issue: I don’t see the Dapr sidecar injected to my pod

Debug

Reference