#!/usr/bin/env bash
# Start any stopped Proxmox containers that back E2E endpoints (502 fixes).
# Uses SSH to Proxmox hosts and pct start for containers that are stopped.
#
# Usage: ./scripts/maintenance/start-stopped-containers-via-ssh.sh [--dry-run]
# Env: PROXMOX_HOST_ML110 (default 192.168.11.10), PROXMOX_HOST_R630_01 (default 192.168.11.11)
#      SSH to root@host; ensure key-based auth or use ssh-agent.
#
# VMIDs: DBIS (10130 frontend, 10150 api primary, 10151 api secondary) + optional 10100,10101,10120;
#        RPC core 2101. Layout: 101xx on ML110 by default (create-dbis-core), 2101 on r630-01.
set -euo pipefail

SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
PROJECT_ROOT="$(cd "$SCRIPT_DIR/../.." && pwd)"
[[ -f "${PROJECT_ROOT}/scripts/lib/load-project-env.sh" ]] && source "${PROJECT_ROOT}/scripts/lib/load-project-env.sh" 2>/dev/null || true

DRY_RUN=false
[[ "${1:-}" == "--dry-run" ]] && DRY_RUN=true

PROXMOX_ML110="${PROXMOX_HOST_ML110:-${PROXMOX_HOST:-192.168.11.10}}"
PROXMOX_R630_01="${PROXMOX_HOST_R630_01:-192.168.11.11}"

# DBIS Core containers (often on ml110)
VMIDS_DBIS="10130 10150 10151 10100 10101 10120"
# RPC core (r630-01 per ALL_VMIDS)
VMIDS_RPC="2101"

log_info() { echo -e "\033[0;34m[INFO]\033[0m $1"; }
log_ok()   { echo -e "\033[0;32m[✓]\033[0m $1"; }
log_warn() { echo -e "\033[0;33m[⚠]\033[0m $1"; }
log_err()  { echo -e "\033[0;31m[✗]\033[0m $1"; }

run_ssh() {
  local host="$1"
  shift
  ssh -o ConnectTimeout=10 -o StrictHostKeyChecking=no root@"$host" "$@"
}

start_stopped_on_host() {
  local host="$1"
  local vmids_list="$2"
  local label="$3"
  log_info "Host $host ($label): checking VMIDs $vmids_list"
  if ! run_ssh "$host" "echo OK" &>/dev/null; then
    log_warn "Cannot SSH to $host; skipping."
    return 0
  fi
  for vmid in $vmids_list; do
    local status
    status=$(run_ssh "$host" "pct status $vmid 2>/dev/null | awk '{print \$2}'" 2>/dev/null || echo "missing")
    if [[ "$status" == "missing" || -z "$status" ]]; then
      log_info "  VMID $vmid: not on this host or unknown, skip"
      continue
    fi
    if [[ "$status" == "running" ]]; then
      log_ok "  VMID $vmid: already running"
      continue
    fi
    if [[ "$DRY_RUN" == true ]]; then
      log_info "  VMID $vmid: would run pct start $vmid (current: $status)"
      continue
    fi
    log_info "  VMID $vmid: starting (was $status)..."
    if run_ssh "$host" "pct start $vmid" 2>/dev/null; then
      log_ok "  VMID $vmid: started"
    else
      log_err "  VMID $vmid: start failed"
    fi
  done
}

echo ""
echo "=== Start stopped Proxmox containers (E2E 502 fix) ==="
echo "  dry-run=$DRY_RUN"
echo ""

start_stopped_on_host "$PROXMOX_ML110" "$VMIDS_DBIS" "ML110 (DBIS)"
echo ""
start_stopped_on_host "$PROXMOX_R630_01" "$VMIDS_RPC" "r630-01 (RPC 2101)"
# Also try DBIS on r630-01 in case layout differs
start_stopped_on_host "$PROXMOX_R630_01" "$VMIDS_DBIS" "r630-01 (DBIS if present)"

echo ""
log_ok "Done. Re-run E2E routing: ./scripts/verify/verify-end-to-end-routing.sh"
echo ""