Kubernetes
Obtain a registration token
Please refer to the Quickstart Guide to obtain a registration token.
Deploy a new Conflux Node
To deploy a new Conflux Node, simply use the following deployment manifest:
apiVersion: apps/v1
kind: Deployment
metadata:
name: conflux
namespace: default # adjust to your namespace
spec:
replicas: 1 # adjust to your desired number of replicas
selector:
matchLabels:
name: conflux
template:
metadata:
labels:
name: conflux
spec:
containers:
- name: conflux
image: veilnet/conflux:beta
imagePullPolicy: Always
securityContext:
privileged: true
env:
- name: VEILNET_REGISTRATION_TOKEN
value: <registration_token> # replace with your registration token
- name: VEILNET_PORTAL
value: "true" # Portal mode allows you to access the entire container network, host, and local network.
resources:
requests:
cpu: "500m" # adjust according to your needs
memory: "512Mi" # adjust according to your needs
limits:
cpu: "500m" # adjust according to your needs
memory: "512Mi" # adjust according to your needs
Enable Horizontal Scaling
To enable horizontal scaling, simply create a horizontal pod autoscaler:
apiVersion: autoscaling/v2
kind: HorizontalPodAutoscaler
metadata:
name: conflux
namespace: default # adjust to your namespace
spec:
scaleTargetRef:
apiVersion: apps/v1
kind: Deployment
name: conflux
minReplicas: 1
maxReplicas: 3
metrics:
- type: Resource
resource:
name: cpu
target:
type: Utilization
averageUtilization: 70
- type: Resource
resource:
name: memory
target:
type: Utilization
averageUtilization: 70
New pods of VeilNet Conflux will be created automatically when the CPU or memory utilization is greater than 70%. You can adjust the metrics as needed, according to the offical kubernetes documentation.
Enable High Availability
When the number of replicas is greater than 1, VeilNet Conflux will automatically be in high availability mode. Each pod could be used as the connector to the container network, host, and local network. The Anchor Protocol will automatically choose the best pod to connect to the destination, and multiple pods could be used concurrently to achieve higher throughput.
There is nothing you need to do to enable high availability.
Warning: Envrionment variables
VEILNET_CONFLUX_CIDRcan be used to set the CIDR of the VeilNet Conflux pod, but this will disable the high availability mode because new instance will attempt to register with the same CIDR, which will cause conflicts.