NSP Kubernetes Platform Communications

Overview

The tables provided in this section identify the listening ports on a deployer node and on worker nodes for an NSP cluster deployment. (Note: worker nodes of an NSP cluster includes MDM only and PCE only nodes.) These ports must be accessible between the deployer and worker nodes within an NSP deployment. The SSH ports on all servers must be accessible by a system administrator for installation and maintenance functions.

Note: The ‘node exporter’ referred to in this section is used by Prometheus to pull performance metrics from a VM.

Table 6-14: Ports used by deployer node

Default port(s)

Type

Application

22

TCP

SSH

111

TCP

rpcbind

443

TCP

HTTPS

6443

TCP

kubernetes API server

8443

TCP

helm repo, container registry

9100

TCP

node exporter

10250

TCP

kubelet metrics

30000-32767

TCP

kube proxy

Table 6-15: Ports used by worker nodes

Default Port(s)

Type

Node(s)

Application

22

TCP

all

sshd

53

TCP

all

node-cache

111

TCP

all

rpcbind

111

UDP

node4, node5, node6

rpcbind

179

TCP

all

bird

443

TCP

any two nodes in enhanced profile, any one node for all other profiles

nsp-piraeus-operator

2049

TCP

one of the nodes that has linstor-controller

NFS server service

2375

TCP

all

containerd

2379

TCP

all

etcd

2380

TCP

all

etcd

3370

TCP

last 3 nodes of NSP cluster

linstor-controller

4789

UDP

all

VXLAN encapsulation

6443

TCP

all

kubernetes API server

7472

TCP

all

metalLB (metrics)

7946

TCP & UDP

all

metalLB (load balancer communications)

8081

TCP

all

nginx

9100

TCP

all

node exporter

9253

TCP

all

node cache

9254

TCP

all

node cache

9353

TCP

all

node-cache

10250

TCP

all

kubelet metrics

10251

TCP

all

kube-scheduler

10256

TCP

all

kube-proxy

10257

TCP

all

kube controller

10259

TCP

all

kube scheduler

30000-32767

TCP

all

kube-proxy