K
Q

how to communicate with daemonset pod from another pod in the same node?

June 12, 2019

i want a daemonset-redis where every node will have it's own caching and each deployment pod will communicate with it's local daemonset-redis how to achieve it? how to reference daemonset pod in the same node from within docker-container?

UPDATE: i rather not use service option and make sure each pod access its local daemonset

apiVersion: apps/v1
kind: DaemonSet
metadata:
  name: redislocal
spec:
  selector:
    matchLabels:
      name: redislocal
  template:
    metadata:
      labels:
        name: redislocal
    spec:
      hostNetwork: true
      containers:
      - name: redislocal
        image: redis:5.0.5-alpine
        ports:
        - containerPort: 6379
          hostPort: 6379
-- Oren Lalezari
kubernetes
daemonset

Similar Questions

Ingress configuration for k8s in different namespaces
The node was low on resource: ephemeral-storage
Unable to connect to the server: net/http: TLS handshake timeout
netstat showing foreign ports as kubernetes:port. What does this mean?
Kubectl command throwing error: Unable to connect to the server: getting credentials: exec: exit status 2

3 Answers

August 2, 2019

There is a way of not using a 

service
.

You can Expose Pod Information to Containers Through Environment Variables.

And you can use 

status.hostIP
to know the ip address of node where 
pod
is running. This was introduced in Kubernetes 1.7 link

You can add that to your 

pod
or 
deployment
yaml:

env:
- name: HOST_IP
  valueFrom:
    fieldRef:
      fieldPath: status.hostIP

It will set an variable 

HOST_IP
which will have a value of node ip on which the 
pod
is running, then you can use it to connect to a local 
DeamonSet
.

-- Crou
Source: StackOverflow
January 2, 2025

This is an old question but I recently had to tackle the exact same problem, and just using the node IP did not really do it when deploying the redis Daemonset through the Bitnami\Redis helm chart. The solution was simpler - use 

internalTrafficPolicy: Local
on the service configuration, which enforces calls made to the clusterIP associated with the service to be redirected locally (and only locally):

The kube-proxy filters the endpoints it routes to based on the spec.internalTrafficPolicy setting. When it's set to Local, only node local endpoints are considered. When it's Cluster (the default), or is not set, Kubernetes considers all endpoints.

source

-- yuvi
Source: StackOverflow
June 12, 2019

you should define a service ( selecting all redis pods ) and then communicate with redis from other pods

-- P Ekambaram
Source: StackOverflow