fix: add clean-data playbook to prevent storage-related deployment failures

[KatyaRyazantseva]

While running devnet-1, one of the client nodes reached its storage limit. A new ansible deployment failed when copying genesis files to the server due to insufficient storage. To restart the devnet, manual cleaning was required. This PR adds centralized data cleaning that runs before genesis generation when using the --cleanData flag.

The clean-data playbook can be called independently for cleaning up node data directories. It supports all deployment modes: site.yml, deploy-nodes.yml, and tag-based execution.

Scenarios tested on Lighthouse server

# Primary use case (via spin-node.sh)
NETWORK_DIR=local-devnet ./spin-node.sh --node lighthouse_0 --deploymentMode ansible --useRoot --cleanData

# Direct ansible with site.yml
./ansible-deploy.sh --node lighthouse_0 --clean-data

 #With deploy-nodes.yml
./ansible-deploy.sh --playbook deploy-nodes.yml --node lighthouse_0 --clean-data

# With tags
./ansible-deploy.sh --node lighthouse_0 --network-dir ansible-devnet --tags lighthouse --clean-data

# Independent cleaning (no deployment)
ansible-playbook -i ansible/inventory/hosts.yml ansible/playbooks/clean-data.yml \
    -e "genesis_dir=$(pwd)/ansible-devnet/genesis" \
    -e "node_names=lighthouse_0"

[ch4r10t33r]

We appear to be double cleaning.
When clean_data=true is passed via site.yml, data gets cleaned:

• First in clean-data.yml (step 1)
• Then again in each role's tasks
  Should we consider using clean-data.yml for pre-deployment cleaning alone and remove role level cleaning when using site.yml?

[ch4r10t33r]

Can we extract this into a shared task file under roles/common/tasks/clean-node-data.yml and use include_tasks in each role.

- name: Clean node data if requested
  include_tasks: "{{ playbook_dir }}/../roles/common/tasks/clean-node-data.yml"
  when: clean_data | default(false) | bool

This will modularize the code (instead of copying the same content across 5 different roles (which is likely to increase in future).

[ch4r10t33r]

The find task to check if a directory has contents before deletion is redundant. Ansible's file: state=absent is idempotent. It will succeed whether the directory exists, is empty, or has contents.

[ch4r10t33r]

This is a nit, but can we please add a check to validate if yq is installed before using it here?