多台服务器批量发布arcgisserver服务并缓存切片

tomelrg2025-08-19 10:23

需求：使用多台服务器发布一批已经建好的mxd工程切片服务。

解决思路：鉴于开发时间和成本考虑，将mxd文件放在共享文件夹中，每台服务器上部署一个发布服务切片的脚本，多台同时访问共享文件夹，当一台访问某个mxd文件时创建一个标识文件，表示正在发布服务或者已发布完成。发布服务超时删除标识文件，其他服务器可以接手。

参数比较多，我已经将参数和代码分离，分成了几个类。下面给出初步的脚本，有兴趣的同学可以用大模型将其优化或者分离。其实也可以写一个前后端的web程序，部署在多台服务器上，解决好任务分配和超时的问题，效果会更好。

python 复制代码 
# -*- coding: utf-8 -*-
import arcpy
import os
import sys
import time
import datetime
import traceback
import codecs
import socket
import shutil

arcpy.env.overwriteOutput = True

# ===================== 日志函数 =====================
def log(msg, run_log_file):
    ts = datetime.datetime.now().strftime('%Y-%m-%d %H:%M:%S')
    line = u"[{0}] {1}".format(ts, msg)
    try:
        print line.encode(sys.stdout.encoding or 'gbk')
    except:
        pass
    with codecs.open(run_log_file, "a", "utf-8") as f:
        f.write(line + u"\r\n")

def log_stats(msg, stats_log_file):
    ts = datetime.datetime.now().strftime('%Y-%m-%d %H:%M:%S')
    with codecs.open(stats_log_file, "a", "utf-8") as f:
        f.write(u"[{0}] {1}\r\n".format(ts, msg))

# ===================== ArcGIS Server =====================
def get_server_connection_path(ags_dir, ags_conn_name):
    conn_path = os.path.join(ags_dir, ags_conn_name)
    if not os.path.exists(conn_path):
        raise IOError(u"找不到 ArcGIS Server 连接文件: {0}".format(conn_path))
    return conn_path

# ===================== 标志文件操作 =====================
def mxd_done_path(mxd_file):
    return mxd_file + ".done"

def mxd_lock_path(mxd_file):
    return mxd_file + ".lock"

def acquire_mxd_task(mxd_file, server_name, lock_timeout, run_log_file):
    lock_file = mxd_lock_path(mxd_file)
    done_file = mxd_done_path(mxd_file)
    if os.path.exists(done_file):
        return False
    if os.path.exists(lock_file):
        last_mod = os.path.getmtime(lock_file)
        if (time.time() - last_mod) > lock_timeout:
            log(u"检测到挂起锁文件已超时，删除锁：{0}".format(lock_file), run_log_file)
            os.remove(lock_file)
        else:
            return False
    with open(lock_file, "w") as f:
        f.write("SERVER={0}\nTIME={1}\n".format(server_name, datetime.datetime.now()))
    return True

def release_mxd_task(mxd_file, server_name, success, run_log_file):
    lock_file = mxd_lock_path(mxd_file)
    done_file = mxd_done_path(mxd_file)
    try:
        if success:
            if os.path.exists(lock_file):
                os.remove(lock_file)
            with open(done_file, "w") as f:
                f.write("DONE={0}\nSERVER={1}\n".format(datetime.datetime.now(), server_name))
            log(u"任务完成，done 文件创建成功: {0}".format(done_file), run_log_file)
        else:
            if os.path.exists(lock_file):
                os.remove(lock_file)
            log(u"任务失败，已删除锁文件，允许其他服务器接手: {0}".format(lock_file), run_log_file)
    except Exception as e:
        log(u"释放锁/创建 done 文件异常: {0}".format(e), run_log_file)

# ===================== 临时 MXD 目录 =====================
def create_temp_mxd(mxd_file, temp_dir_root, server_name):
    temp_server_dir = os.path.join(temp_dir_root, server_name)
    if not os.path.exists(temp_server_dir):
        os.makedirs(temp_server_dir)
    temp_mxd = os.path.join(temp_server_dir, os.path.basename(mxd_file))
    shutil.copy2(mxd_file, temp_mxd)
    return temp_mxd

# ===================== 强校验切片完成 =====================
def check_jobstatus_gdb(cache_path, run_log_file):
    try:
        jobstatus_gdb = os.path.join(cache_path, "Layers", "Status.gdb")
        if not os.path.exists(jobstatus_gdb):
            log(u"未找到 JobStatus.gdb: {0}".format(jobstatus_gdb), run_log_file)
            return False
        arcpy.env.workspace = jobstatus_gdb
        fcs = arcpy.ListFeatureClasses("*JobStatus*")
        if not fcs:
            log(u"未找到 JobStatus 表", run_log_file)
            return False
        table = fcs[0]
        with arcpy.da.SearchCursor(table, ["TotalTasks", "DoneTasks"]) as cursor:
            for row in cursor:
                total, done = row
                if total != done:
                    return False
        return True
    except Exception as e:
        log(u"check_jobstatus_gdb 异常: {0}".format(e), run_log_file)
        return False

def wait_for_cache_completion(result, service_name, cache_root, service_dir, run_log_file):
    cache_path = os.path.join(cache_root, service_dir + "_" + service_name)
    try:
        log(u"等待切片任务完成...", run_log_file)
        while True:
            if result.status == 4 and check_jobstatus_gdb(cache_path, run_log_file):
                log(u"切片任务已完全完成 ✅", run_log_file)
                break
            else:
                log(u"切片未完成，等待中...", run_log_file)
                time.sleep(10)
    except Exception as e:
        log(u"切片任务监控异常: {0}".format(e), run_log_file)

# ===================== 发布与切片 =====================
def publish_and_cache(mxd_file, server_conn_path, cache_root, service_dir,
                      dpi, tile_size, scales, scales2, cache_tile_format,
                      tile_compression, thread_count, run_log_file):
    service_name = os.path.splitext(os.path.basename(mxd_file))[0]
    service_path = os.path.join(server_conn_path[:-4], service_dir, service_name + ".MapServer")
    try:
        log(u"开始发布 MXD: {0}".format(mxd_file), run_log_file)
        temp_mxd = create_temp_mxd(mxd_file, temp_dir_root=os.path.dirname(os.path.dirname(mxd_file)), server_name=SERVER_NAME)

        sddraft = temp_mxd.replace(".mxd", ".sddraft")
        sd = temp_mxd.replace(".mxd", ".sd")

        mxd_doc = arcpy.mapping.MapDocument(temp_mxd)
        arcpy.mapping.CreateMapSDDraft(
            mxd_doc, sddraft, service_name, "ARCGIS_SERVER",
            server_conn_path, True, None, None, service_dir
        )
        del mxd_doc

        analysis = arcpy.mapping.AnalyzeForSD(sddraft)
        if analysis['errors'] != {}:
            log(u"服务草稿存在错误: {0}".format(analysis['errors']), run_log_file)
            raise RuntimeError("AnalyzeForSD errors")

        arcpy.StageService_server(sddraft, sd)
        arcpy.UploadServiceDefinition_server(
            sd, server_conn_path, "", "", "EXISTING", service_dir
        )
        log(u"服务发布完成: {0}".format(service_name), run_log_file)

        log(u"创建缓存 schema: {0}".format(service_name), run_log_file)
        arcpy.CreateMapServerCache_server(
            service_path, cache_root, "NEW", "CUSTOM",
            18, dpi, tile_size,
            "", "-180 90", scales, cache_tile_format, tile_compression, "COMPACT"
        )

        log(u"开始生成切片: {0}".format(service_name), run_log_file)
        result = arcpy.ManageMapServerCacheTiles_server(
            service_path, scales2, "RECREATE_ALL_TILES", thread_count, "", "", "WAIT"
        )
        wait_for_cache_completion(result, service_name, cache_root, service_dir, run_log_file)

        os.remove(temp_mxd)
        log(u"临时 MXD 删除: {0}".format(temp_mxd), run_log_file)

    except Exception as e:
        log(u"发布/切片出错: {0}\r\n{1}".format(e, traceback.format_exc()), run_log_file)
        raise

# ===================== 主流程 =====================
def main():
    # 配置参数（全部局部变量）
    mxd_dir = r"D:\test\mxd"
    temp_dir_root = r"D:\test\temp_publish"
    ags_dir = os.path.join(os.environ['APPDATA'], r"ESRI\Desktop10.2\ArcCatalog")
    ags_conn_name = "arcgis on localhost_6080 (admin).ags"
    service_dir = "zw"
    cache_dir = r"D:\arcgisserver\directories\arcgiscache"
    cache_root = os.path.join(cache_dir, service_dir)

    server_name = os.environ.get("COMPUTERNAME") or socket.gethostname()

    log_dir = r"D:\test\log"
    if not os.path.exists(log_dir):
        os.makedirs(log_dir)
    run_log_file = os.path.join(log_dir,
        "publish_log_{0}.log".format(datetime.datetime.now().strftime('%Y%m%d_%H%M%S')))
    stats_log_file = os.path.join(log_dir,
        "stats_log_{0}.log".format(datetime.datetime.now().strftime('%Y%m%d')))

    scales = "591658710.9;295829355.45;147914677.725;73957338.8625;36978669.43125;18489334.715625;9244667.3578125;4622333.67890625;2311166.839453125;1155583.4197265625;577791.7098632812;288895.8549316406;144447.9274658203;72223.96373291015;36111.98186645508;18055.99093322754;9027.99546661377"
    scales2 = "72223.96373291015;36111.98186645508;18055.99093322754;9027.99546661377"

    cache_tile_format = "MIXED"
    tile_compression = 75
    tile_size = "256 x 256"
    dpi = 96
    thread_count = 24
    lock_timeout = 24 * 60 * 60  # 24 小时

    try:
        conn_path = get_server_connection_path(ags_dir, ags_conn_name)
        log(u"本机: {0}".format(server_name), run_log_file)
        log(u"服务文件夹: {0}".format(service_dir), run_log_file)

        mxd_list = [os.path.join(mxd_dir, f) for f in os.listdir(mxd_dir) if f.lower().endswith(".mxd")]
        if not mxd_list:
            log(u"未找到 MXD 文件", run_log_file)
            sys.exit(0)

        for mxd_file in mxd_list:
            svc_name = os.path.splitext(os.path.basename(mxd_file))[0]
            try:
                if not acquire_mxd_task(mxd_file, server_name, lock_timeout, run_log_file):
                    continue
                publish_and_cache(mxd_file, conn_path, cache_root, service_dir,
                                  dpi, tile_size, scales, scales2, cache_tile_format,
                                  tile_compression, thread_count, run_log_file)
                release_mxd_task(mxd_file, server_name, success=True, run_log_file=run_log_file)
                log_stats(u"完成: {0}".format(svc_name), stats_log_file)
            except Exception as e:
                release_mxd_task(mxd_file, server_name, success=False, run_log_file=run_log_file)
                log_stats(u"失败: {0} 错误: {1}".format(svc_name, e), stats_log_file)

        log(u"全部 MXD 发布与切片任务完成！", run_log_file)

    except Exception as e:
        log(u"主程序异常: {0}\r\n{1}".format(e, traceback.format_exc()), run_log_file)

if __name__ == "__main__":
    main()
相关推荐
R-G-B39 分钟前
OpenCV Python——Numpy基本操作（Numpy 矩阵操作、Numpy 矩阵的检索与赋值、Numpy 操作ROI）
python·opencv·numpy·numpy基本操作·numpy 矩阵操作·numpy 矩阵的检索与赋值·numpy 操作roi
细节处有神明1 小时前
Jupyter 中实现交互式图表：ipywidgets 从入门到部署
ide·python·jupyter
小小码农一只1 小时前
Python 爬虫实战：玩转 Playwright 跨浏览器自动化（Chromium/Firefox/WebKit 全支持）
爬虫·python·自动化
赏点剩饭7782 小时前
linux中的hostpath卷、nfs卷以及静态持久卷的区别
linux·运维·服务器
深盾安全2 小时前
Python脚本安全防护策略全解析（上）
python
杜子不疼.2 小时前
《Python学习之使用标准库：从入门到实战》
开发语言·python·学习
神鸟云2 小时前
DELL服务器 R系列 IPMI的配置
linux·运维·服务器·网络·边缘计算·pcdn
胡耀超2 小时前
从哲学（业务）视角看待数据挖掘：从认知到实践的螺旋上升
人工智能·python·数据挖掘·大模型·特征工程·crisp-dm螺旋认知·批判性思维
herderl2 小时前
**僵尸进程（Zombie Process）** 和**孤儿进程（Orphan Process）**
linux·运维·服务器·网络·网络协议
热门推荐
01UV安装并设置国内源02Qwen3-Coder 快速上手教程 | Qwen Code + Claude Code03【2025.08.06最新版】Android Studio下载、安装及配置记录（自动下载sdk）04KGG转MP3工具|非KGM文件|解密音频05TRAE 规则（Rules）配置指南：个人习惯、团队规范与最佳实践062025最新国内服务器可用docker源仓库地址大全（2025年8月更新）07NVIDIA显卡驱动、CUDA、cuDNN 和 TensorRT 版本匹配指南08TRAE Rules 实践：为项目配置 6A 工作流09蜘蛛磁力 搜索引擎大全，如何使用蜘蛛磁力查找磁力链接10阿里开源首个图像生成基础模型——Qwen-Image本地部署教程，超强中文渲染能力刷新SOTA！