chore: fix underlying causes of playbook warnings

[Q-Dub]

Summary

Fix Ansible playbook warnings by addressing root causes, not suppressing them.

• Export inventory: Document running exported playbooks from bloom-playbook/ so ansible.cfg loads inventory.ini (export wiring already on main)
• Sudo preflight: Replace ignore_errors + registered-result check with block/rescue; fail immediately with actionable message when privilege escalation is unavailable
• Kernel modules: Skip modprobe when iscsi_tcp / dm_mod are already loaded (built-in or present under /sys/module)
• Inotify sysctl: Pass string value to the sysctl module
• Logrotate cron task: Remove duplicate enabled YAML key
• Local-path waits: Use kubectl wait --for=create before readiness checks instead of Ansible retry loops that emit FAILED - RETRYING

Non-goals / rejected approaches

• warn: false on RKE2 curl | sh install tasks — removed; unsupported on current Ansible
• ignore_errors: true on sudo preflight — removed in favor of explicit block/rescue failure

Test plan

• ./bloom cli bloom.yaml --export — confirm inventory.ini and ansible.cfg present
• cd bloom-playbook && ansible-playbook cluster-bloom.yaml — no inventory warnings
• sudo ./bloom cli bloom.yaml — sudo preflight still fails clearly without passwordless sudo
• QEMU CI integration test passes
• RKE2 prep: no modprobe warnings when dm_mod is built-in
• Local-path deploy: no FAILED - RETRYING lines during namespace/deployment waits

[pre]

Motivation for silencing the warnings and ignoring errors is not clear to me 🤔

[Q-Dub]

To clarify: the scope is to resolve underlying issues which are triggering warnings, not to silence anything.