refactor: rename failover/HA → fallback across entire codebase

Removes all references to "failover" and "HA" (high availability)
terminology from variable names, config keys, state values, rsync
profile names, directory paths, and user-visible strings.

Mapping:
  FAILOVER_*              → FALLBACK_*
  FAILOVER_HOST*_RUNS_FOR → FALLBACK_HOST*_COVERS
  critical-failover       → critical-fallback
  emby-failover           → emby-fallback
  appdata-Failover/       → appdata-Fallback/
  "FAILOVER" state value  → "FALLBACK"
  failover_start key      → fallback_start
  Failover/ directory     → Fallback/
  failover.sh             → fallback.sh
  failover_state.db       → fallback_state.db
  -Failover folder suffix → -Fallback

State machine: NORMAL | FALLBACK | DARK (unchanged)

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

Orchestrators/array_start.sh

@@ -20,7 +20,7 @@
 #   CONTINUOUS (run until array stops):
 #     unRAID_Essentials/system_watchdog.sh      — system health monitor (last line of defense)
 #     Docker_Essentials/docker_watchdog.sh      — container health monitor
-#     Failover/failover.sh                      — mutual failover monitor
+#     Fallback/fallback.sh                      — mutual failover monitor
 #
 # ── WHY ORDER MATTERS ─────────────────────────────────────────────────────────────────────────
 #   inotify_tuning.sh     — must run BEFORE Code-Server and other containers start
@@ -30,7 +30,7 @@
 #   docker_network_connect — must run BEFORE watchdogs check container states
 #   system_watchdog.sh    — before docker_watchdog (system > container priority)
 #   docker_watchdog.sh    — before failover (containers must be healthy for failover)
-#   failover.sh           — last — needs everything else stable to make decisions
+#   fallback.sh           — last — needs everything else stable to make decisions
 #
 # ── ONE-SHOT vs CONTINUOUS DETECTION ─────────────────────────────────────────────────────────
 # Script is launched in background with bash script.sh &