Error to setup sentinnel and replicate data between nodes

[julioo]

Hello,

Pods are setup on my GCP cluster but sentinel returns an error and replication between nodes is not ok.

If I try to determine which server the Redis sentinel believes is the master, I get the following :
$ kubectl exec redis-2 -c redis -- redis-cli -p 26379 sentinel get-master-addr-by-name redis
Could not connect to Redis at 127.0.0.1:26379: Connection refused

If I check sentinel logs
$ kubectl logs statefulsets/redis sentinel
Waiting for server
ping: bad address 'redis-0.redis'

And if I set foo 10 on redis-0 I don't get key on redis-2

Why I get bad address ?
What can I do ?
Thank you

[julioo]

Ok almost done, just remove .redis of the hostname, using just redis-0 it works fine on redis-0 pod but
redis-0 is not seen since redis-1 and redis-2

Error from redis on redis-1 and redis-2
5:S 23 Oct 05:26:02.099 * Connecting to MASTER redis-0:6379
5:S 23 Oct 05:26:02.110 # Unable to connect to MASTER: Resource temporarily unavailable

Or from sentinel on redis-1 and redis-2
ping: bad address 'redis-0'
Waiting for server

How to make them communicate ?

[hasonhai]

One of the error may be you forgot to set "spec.clusterIP: None" in the service file.

[AlexProfi]

One of the error may be you forgot to set "spec.clusterIP: None" in the service file.

Did you mean redis-sentinel service file from examples?

apiVersion: v1
kind: Service
metadata:
  labels:
    name: sentinel
    role: service
  name: redis-sentinel
spec:
  ports:
    - port: 26379
      targetPort: 26379
  selector:
    redis-sentinel: "true"

I use sentinel in nginx lua to connect to master for set data. If it will have not the ip, it becomes impossible