Fix Redis cluster hostname announcement using FQDN

[fonsecas72]

Redis cluster nodes were announcing themselves with short hostnames (e.g., my-redis-leader-0) instead of FQDNs, preventing proper cluster communication.

Changed POD_HOSTNAME=$(hostname) to POD_HOSTNAME=$(hostname -f) in the Redis entrypoint script to use the fully qualified domain name (e.g., my-redis-leader-0.my-redis-leader-headless.my-redis-namespace.svc.cluster.local).

Redis cluster nodes can now discover and communicate with each other regardless of namespace

[drivebyer]

It fixed by OT-CONTAINER-KIT/redis-operator#1489 which use init-config init container

[fonsecas72]

thanks @drivebyer it wasn't clear for me. I'll give the last version a try

[fonsecas72]

Getting this with the default redis v7.0.15

  1:M 11 Sep 2025 12:57:40.700 # Server initialized
  1:M 11 Sep 2025 12:57:40.701 * Reading RDB base file on AOF loading...
  1:M 11 Sep 2025 12:57:40.701 # Can't handle RDB format version 12
  1:M 11 Sep 2025 12:57:40.701 # Error reading the RDB base file appendonly.aof.5.base.rdb, AOF loading aborted
  1:M 11 Sep 2025 12:57:40.701 # Unrecoverable error reading the append only file appendonly.aof.5.base.rdb: Invalid argument

After changing to redis v7.4.5...

i'm getting

*** FATAL CONFIG FILE ERROR (Redis 7.4.5) ***
Reading the configuration file, at line 21
>>> 'cluster-announce-ip'
wrong number of arguments

At first (sometimes) and then it fixes by itself. Same behaviour i was having with this PR and why i put the loop in it.

Do you have any suggestion on how to fix it? Maybe some param i might not be setting?
@drivebyer