Files
smart-inspection/scripts/prod-backend.sh
T

260 lines
9.0 KiB
Bash
Executable File

#!/bin/bash
# ============================================================
# Robot AGV 全量启动脚本 v5.0
# 修复:
# - v5.0: 使用公共库重构,减少代码重复
# - v4.0: 彻底杀死 ros2 daemon 进程 + 启动前进程数量检查
# - v3.0: 彻底清理 FastRTPS 共享内存文件(永久修复 DDS 通信问题)
# - v2.7: 添加 ROS_DOMAIN_ID 环境变量传递
# - v2.6: 清理 scan_fixer lock 文件防残留
# ============================================================
set -euo pipefail
# 加载公共库
SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
source "$SCRIPT_DIR/ros-common.sh"
# ============================================================================
# 主流程
# ============================================================================
section "Robot AGV 全量启动 v5.0"
# ============================================================================
# 1. 清理旧环境
# ============================================================================
step "1/8" "清理旧进程和共享内存"
kill_all_soft
kill_all_hard
stop_ros2_daemon
cleanup_fastrtps
# 验证进程已停止
echo " 验证进程停止..."
PROC_COUNT=$(count_residual_processes)
echo " 残留进程数: $PROC_COUNT"
if [ "$PROC_COUNT" -gt 0 ]; then
echo " [WARN] 仍有进程残留,再次强制终止..."
kill_all_hard
PROC_COUNT=$(count_residual_processes)
echo " 清理后残留: $PROC_COUNT"
fi
info "ok" "清理完成"
# ============================================================================
# 2. 启动 ros2 daemon
# ============================================================================
step "2/8" "启动 ros2 daemon"
rm -rf "$FASTRTPS_SHM_DIR"/fastrtps_* 2>/dev/null || true
start_ros2_daemon || true
# ============================================================================
# 3. 启动 bringup (含激光雷达)
# ============================================================================
step "3/8" "启动 AGV Bringup"
cd "$AGV_ROS2_DIR"
rm -rf "$FASTRTPS_SHM_DIR"/fastrtps_* 2>/dev/null || true
nohup bash -c '
source "$1" || exit 1
source "$2" || exit 1
export ROS_DOMAIN_ID="$3"
cd "$4" || exit 1
ros2 launch agv_pro_bringup agv_pro_bringup.launch.py port_name:="$5"
' _ "$ROS_SETUP" "$ROS_WORKSPACE_SETUP" "$ROS_DOMAIN_ID" "$AGV_ROS2_DIR" "$AGV_CONTROLLER_DEVICE" \
> "$BRINGUP_LOG" 2>&1 &
BRINGUP_PID=$!
echo " bringup PID: $BRINGUP_PID"
# 等待 /odom 话题
wait_for_topic /odom 40 || show_log_tail "$BRINGUP_LOG"
# ============================================================================
# 4. 启动系统时钟发布器
# ============================================================================
step "4/8" "启动系统时钟发布器 (clock_publisher)"
nohup bash -c "source \"$ROS_SETUP\" && \
ROS_DOMAIN_ID=$ROS_DOMAIN_ID python3 \"$SCAN_FIXER_DIR/clock_publisher.py\"" \
> "$CLOCK_LOG" 2>&1 &
CLOCK_PID=$!
echo " clock_publisher PID: $CLOCK_PID"
sleep 2
wait_for_topic /clock 10 || show_log_tail "$CLOCK_LOG"
# ============================================================================
# 5. 启动激光时间戳修正节点
# ============================================================================
step "5/8" "启动激光时间戳修正节点"
# 先等待 /scan 话题
if ! wait_for_topic /scan 20; then
echo " [WARN] /scan 未上线,检查 bringup 日志"
fi
nohup bash -c "source \"$ROS_SETUP\" && \
ROS_DOMAIN_ID=$ROS_DOMAIN_ID python3 \"$SCAN_FIXER_DIR/fix_scan_timestamp_v6.py\"" \
> "$SCAN_FIXER_LOG" 2>&1 &
FIXER_PID=$!
echo " fix_scan_timestamp PID: $FIXER_PID"
sleep 5
# 检查是否有多个 fixer 进程
FIXER_COUNT=$(count_matching_processes "fix_scan_timestamp")
if [ "$FIXER_COUNT" -gt 1 ]; then
echo " [WARN] 发现 $FIXER_COUNT 个 fixer 进程,重启..."
pkill -f "fix_scan_timestamp" 2>/dev/null || true
pkill -f "clock_publisher" 2>/dev/null || true
sleep 2
rm -f /tmp/scan_fixer.lock
nohup bash -c "source \"$ROS_SETUP\" && \
ROS_DOMAIN_ID=$ROS_DOMAIN_ID python3 \"$SCAN_FIXER_DIR/fix_scan_timestamp_v6.py\"" \
> "$SCAN_FIXER_LOG" 2>&1 &
FIXER_PID=$!
sleep 3
fi
wait_for_topic /scan_corrected 15 || show_log_tail "$SCAN_FIXER_LOG"
# ============================================================================
# 6. 启动 Nav2
# ============================================================================
step "6/8" "启动 Nav2 导航"
nohup bash -c "source \"$ROS_SETUP\" && source \"$ROS_WORKSPACE_SETUP\" && \
export ROS_DOMAIN_ID=$ROS_DOMAIN_ID && \
ros2 launch agv_pro_navigation2 navigation2_active.launch.py autostart:=True use_rviz:=False" \
> "$NAV2_LOG" 2>&1 &
NAV2_PID=$!
echo " Nav2 PID: $NAV2_PID"
sleep 12
echo " 等待 Nav2 节点就绪..."
wait_for_nodes 'lifecycle_manager_navigation|bt_navigator|controller_server' 3 45 || true
# ============================================================================
# 7. 设置精度参数
# ============================================================================
step "7/8" "设置导航精度参数 (xy_goal_tolerance=0.05m)"
for NODE in /controller_server /bt_navigator /planner_server; do
ros2_exec timeout 1 ros2 param set $NODE general_goal_checker.xy_goal_tolerance 0.05 2>/dev/null || true
ros2_exec timeout 1 ros2 param set $NODE general_goal_checker.yaw_goal_tolerance 0.05 2>/dev/null || true
done
ros2_exec timeout 1 ros2 param set /controller_server FollowPath.xy_goal_tolerance 0.05 2>/dev/null || true
ros2_exec timeout 1 ros2 param set /controller_server general_goal_checker.stateful True 2>/dev/null || true
ros2_exec timeout 1 ros2 param set /controller_server FollowPath.stateful True 2>/dev/null || true
info "ok" "精度参数已设置"
# ============================================================================
# 8. 启动 Flask API
# ============================================================================
step "8/8" "启动 Flask API"
cd "$AGV_APP_DIR"
UV_EXECUTABLE=$(resolve_uv_bin || true)
FLASK_PID=""
if [ -z "$UV_EXECUTABLE" ]; then
echo " [ERROR] 未找到 uv,可设置 UV_BIN=/path/to/uv"
else
nohup bash -c '
source "$1" || exit 1
source "$2" || exit 1
export ROS_DOMAIN_ID="$3"
cd "$4" || exit 1
exec "$5" run --locked python app.py
' _ "$ROS_SETUP" "$ROS_WORKSPACE_SETUP" "$ROS_DOMAIN_ID" "$AGV_APP_DIR" "$UV_EXECUTABLE" \
> "$FLASK_LOG" 2>&1 &
FLASK_PID=$!
echo " Flask PID: $FLASK_PID"
fi
sleep 4
# ============================================================================
# 9. 最终验证
# ============================================================================
section "系统全面验证"
# 验证话题数量
echo ""
echo "验证 ros2 topic list..."
TOPIC_COUNT=$(ros2_topic_count 5)
echo " 话题数量: $TOPIC_COUNT"
if [ "$TOPIC_COUNT" -gt 10 ]; then
info "ok" "ros2 daemon 正常 (${TOPIC_COUNT} 个话题)"
else
info "err" "ros2 topic list 异常 (${TOPIC_COUNT} 个话题,可能 DDS 有问题)"
echo " 手动执行: rm -rf \"$FASTRTPS_SHM_DIR\"/fastrtps_* && ros2 daemon stop && ros2 daemon start"
fi
# 验证关键话题
echo ""
echo "验证关键话题..."
for TOPIC in /odom /scan /cmd_vel /tf /clock /scan_corrected; do
if topic_exists "$TOPIC"; then
info "ok" "$TOPIC"
else
info "warn" "$TOPIC 未找到"
fi
done
# 验证进程数量
echo ""
echo "验证进程数量..."
BRINGUP_PROCS=$(count_matching_processes 'agv_pro_node|lslidar_driver_node')
echo " AGV 核心进程: $BRINGUP_PROCS (应为 2)"
if [ "$BRINGUP_PROCS" -eq 2 ]; then
info "ok" "进程数量正常(无重复)"
elif [ "$BRINGUP_PROCS" -gt 2 ]; then
info "warn" "发现 $BRINGUP_PROCS 个核心进程(可能有残留),建议重启"
else
info "warn" "进程数量异常"
fi
# FastRTPS 共享内存状态
echo ""
echo "FastRTPS 共享内存状态:"
FASTRTPS_NEW=$(count_fastrtps_files)
echo " 当前文件数: $FASTRTPS_NEW (正常运行时会有一些)"
# Flask API 状态
echo ""
echo "验证 Flask API..."
if pgrep -f "app.py" >/dev/null 2>&1; then
info "ok" "Flask 进程运行中"
else
info "err" "Flask 未运行"
fi
# ============================================================================
# 完成
# ============================================================================
section "[OK] 启动完成"
echo ""
echo " 进程状态:"
for proc_info in "bringup:$BRINGUP_PID" "Nav2:$NAV2_PID" "fixer:$FIXER_PID" "Flask:$FLASK_PID"; do
name="${proc_info%%:*}"
pid="${proc_info##*:}"
if [ -n "$pid" ] && ps -p "$pid" >/dev/null 2>&1; then
echo " $name : 运行中 (PID: $pid)"
else
echo " $name : 已退出"
fi
done
echo ""
echo " 日志文件:"
echo " bringup : $BRINGUP_LOG"
echo " Nav2 : $NAV2_LOG"
echo " fixer : $SCAN_FIXER_LOG"
echo " Flask : $FLASK_LOG"
echo ""
echo " 如果仍有问题,请依次执行:"
echo " 1. ./scripts/stop_all.sh"
echo " 2. rm -rf \"$FASTRTPS_SHM_DIR\"/fastrtps_*"
echo " 3. ./scripts/prod-backend.sh"
echo ""