#!/usr/bin/env bash
# Bring the lab back up after a host reboot. Assumes the VMs, network, and
# storage pools were already created by a prior 00-07 run — this does NOT
# (re)install anything, it just re-establishes libvirt/networking state that
# doesn't survive a reboot (modular daemon sockets, the NAT/forward iptables
# rules) and starts the existing VMs.
set -euo pipefail
source "$(dirname "$0")/lab.env"

sudo -v   # prompt for the sudo password once, up front

# Re-run the (idempotent) daemon + pool + network setup. Safe to do before any
# VM is started: 01-network.sh recreates the pve-net bridge, which would orphan
# any already-running VM's tap device — so this must happen first, while all
# VMs are still shut off.
echo "==> Re-establishing libvirt daemons, storage pools, and pve-net (with NAT rules)"
"${SCRIPT_DIR}/00-prereqs.sh"
"${SCRIPT_DIR}/01-network.sh"

echo "==> Starting VMs"
missing=0
for name in "$PVE1_NAME" "$PVE2_NAME" "$STORE_NAME"; do
  state=$(sudo virsh domstate "$name" 2>/dev/null || echo "missing")
  case "$state" in
    running)   echo "    $name already running" ;;
    "shut off") echo "    starting $name"; sudo virsh start "$name" ;;
    paused)    echo "    resuming $name"; sudo virsh resume "$name" ;;
    missing)   echo "    ERROR: $name is not defined in libvirt"; missing=1 ;;
    *)         echo "    $name is '$state'; leaving as-is" ;;
  esac
done
if [[ "$missing" == 1 ]]; then
  echo "One or more VMs don't exist yet — run the full ./00 through ./07 build first."
  exit 1
fi

wait_ssh() {
  local ip="$1" name="$2" tries=0
  echo "==> Waiting for $name ($ip) over SSH"
  while ! lab_ssh "$ip" true >/dev/null 2>&1; do
    sleep 5
    tries=$((tries+1))
    if (( tries % 12 == 0 )); then echo "    still waiting for $name ... (${tries}0s)"; fi
    if (( tries > 60 )); then echo "    WARN: $name not reachable after 5m; check 'virsh console $name'"; return 1; fi
  done
  echo "    $name is up"
}
wait_ssh "$PVE1_IP" "$PVE1_NAME" || true
wait_ssh "$PVE2_IP" "$PVE2_NAME" || true
wait_ssh "$STORE_IP" "$STORE_NAME" || true

echo "==> Cluster status"
lab_ssh "$PVE1_IP" "pvecm status" 2>/dev/null || echo "    (pve1 not reachable yet; re-run this script or check the VM console)"

echo
echo "Lab is up. Web UIs: https://${PVE1_IP}:8006  https://${PVE2_IP}:8006"
echo "Run ./cluster-info.sh to refresh CLUSTER-INFO.md, or ./verify.sh for a full check."