Optimize Kafka topic initialization on re-deploy

[ehelms]

Why are you introducing these changes? (Problem description, related links)

On every re-deploy, the Kafka topic initialization task (iop_kafka : Initialize Kafka topics) runs the expensive init.sh --create operation, which spawns 17 parallel JVMs to create topics. This takes ~11.4 seconds even when all topics already exist and no actual work is needed.

This pattern mirrors the puppet-iop implementation (puppet-iop/manifests/core_kafka.pp:80-91), which guards topic creation with an unless check.

What are the changes introduced in this pull request?

• Add Check whether Kafka topics already exist task that runs /opt/kafka/init.sh --check before topic creation
• Make the Initialize Kafka topics task conditional on iop_kafka_check.rc != 0
• The --check operation runs a single JVM vs 17 parallel JVMs in --create

How to test this pull request

Steps to reproduce:

• Deploy with IoP enabled: ./foremanctl deploy --flavor katello --features iop
• Verify topics are created: podman exec iop-core-kafka /opt/kafka/init.sh --check (should succeed)
• Re-deploy without changes: ./foremanctl deploy --flavor katello --features iop
• Observe timing: iop_kafka : Check whether Kafka topics already exist should run in ~1-2s and Initialize Kafka topics should be skipped
• Compare to before: previously Initialize Kafka topics ran every deploy and took ~11.4s

Checklist

• Tests added/updated (if applicable)
• Documentation updated (if applicable)