fix: remove stale wg0 interface before container start

- deploy.sh: add 'ip link del wg0' cleanup in force_stop() for both machines
- deploy.sh: add wg0 cleanup before container start (idempotent re-run)
- Fixes 'wg-quick: wg0 already exists' crash on redeploy

deploy.sh

@@ -337,6 +337,9 @@ start_container() {
 
     header "Starting ${role} container: ${host}"
 
+    # Remove stale wg0 interface (leftover from previous container with host network)
+    ssh_remote "$host" "$pass" "ip link del wg0 2>/dev/null || true"
+
     ssh_remote "$host" "$pass" "
         cd ${target}
         docker compose -f ${compose_file} up -d --remove-orphans 2>&1
@@ -349,19 +352,21 @@ start_container() {
 force_stop() {
     header "Force-stopping existing containers"
 
-    info "Stopping VPS container..."
+    info "Stopping VPS container and cleaning up..."
     ssh_remote "$VPS_HOST" "$VPS_PASS" "
         cd /opt/wireguard-vps-tunnel 2>/dev/null || exit 0
         docker compose -f docker-compose.server.yml down --remove-orphans 2>/dev/null || true
+        ip link del wg0 2>/dev/null || true
     " || true
 
-    info "Stopping client container..."
+    info "Stopping client container and cleaning up..."
     ssh_remote "$CLIENT_HOST" "$CLIENT_PASS" "
         cd /opt/wireguard-vps-tunnel 2>/dev/null || exit 0
         docker compose -f docker-compose.client.yml down --remove-orphans 2>/dev/null || true
+        ip link del wg0 2>/dev/null || true
     " || true
 
-    success "Existing containers stopped"
+    success "Existing containers stopped and interfaces cleaned"
 }
 
 # ── Wait for healthy containers ─────────────────────────────────────────────