Question

在Program.cs中，我有以下方法检查和同步5 SQL数据库与中央服务器。每一个都与另一个是分开的，所以我想通过让它们全部同时运行来加快程序的加载时间。

不幸的是，它是非常片状工作一次而不是下一次。本地数据库是SQLExpress 2005，中央数据库是SQL Server Standard 2005。

这两种连接的数量是否有限制？背景工人怎么样，我可以一次只运行这么多吗？我相信有更多雄辩的方法可以做到这一点，我很乐意听到（见）他们。

这就是我在Program.cs中的Main（）中调用它的方法 - ＆gt;

if(IsSqlAvailable()) SyncNow();

internal static void SyncNow()
    {



            #region ConnectDB Merge Sync Background Thread

            BackgroundWorker connectBW = new BackgroundWorker
                                             {
                                                 WorkerReportsProgress = false,
                                                 WorkerSupportsCancellation = true
                                             };
            connectBW.DoWork += new DoWorkEventHandler(connectBW_DoWork);
            if (connectBW.IsBusy != true)
                connectBW.RunWorkerAsync();

            #endregion

            #region aspnetDB Merge Sync Background Thread

            BackgroundWorker aspBW = new BackgroundWorker
                                         {
                                             WorkerReportsProgress = false,
                                             WorkerSupportsCancellation = true
                                         };
            aspBW.DoWork += new DoWorkEventHandler(aspBW_DoWork);
            if (aspBW.IsBusy != true)
                aspBW.RunWorkerAsync();

            #endregion

            #region MatrixDB Merge Sync Background Thread

            BackgroundWorker matrixBW = new BackgroundWorker
                                            {
                                                WorkerReportsProgress = false,
                                                WorkerSupportsCancellation = true
                                            };
            matrixBW.DoWork += new DoWorkEventHandler(matrixBW_DoWork);
            if (matrixBW.IsBusy != true)
                matrixBW.RunWorkerAsync();

            #endregion



            #region CMODB Merge Sync Background Thread

            BackgroundWorker cmoBW = new BackgroundWorker
                                         {
                                             WorkerReportsProgress = false,
                                             WorkerSupportsCancellation = true
                                         };
            cmoBW.DoWork += new DoWorkEventHandler(cmoBW_DoWork);
            if (cmoBW.IsBusy != true)
                cmoBW.RunWorkerAsync();

            #endregion

            #region MemberCenteredPlanDB Merge Sync Background Thread

            BackgroundWorker mcpBW = new BackgroundWorker
                                         {
                                             WorkerReportsProgress = false,
                                             WorkerSupportsCancellation = true
                                         };
            mcpBW.DoWork += new DoWorkEventHandler(mcpBW_DoWork);
            if (mcpBW.IsBusy != true)
                mcpBW.RunWorkerAsync();

            #endregion

    }

    static void mcpBW_DoWork(object sender, DoWorkEventArgs e)
    {
        BackgroundWorker worker = sender as BackgroundWorker;
        try
        {
            MergeRepl mcpMergeRepl = new MergeRepl(SystemInformation.ComputerName + "\\SQLEXPRESS", "WWCSTAGE", "MemberCenteredPlan", "MemberCenteredPlan", "MemberCenteredPlan");
            mcpMergeRepl.RunDataSync();
            areAllInSync += 1; 
        }
        catch (Exception)
        {
            if (worker != null) worker.CancelAsync();
        }
    }

    static void cmoBW_DoWork(object sender, DoWorkEventArgs e)
    {
        BackgroundWorker worker = sender as BackgroundWorker;
        try
        {
            MergeRepl cmoMergeRepl = new MergeRepl(SystemInformation.ComputerName + "\\SQLEXPRESS", "WWCSTAGE", "CMO", "CMO", "CMO");
            cmoMergeRepl.RunDataSync();
            areAllInSync += 1; 
        }
        catch (Exception)
        {
            if (worker != null) worker.CancelAsync();
        }
    }

    static void connectBW_DoWork(object sender, DoWorkEventArgs e)
    {
        BackgroundWorker worker = sender as BackgroundWorker;
        try
        {
            MergeRepl connectMergeRepl = new MergeRepl(SystemInformation.ComputerName + "\\SQLEXPRESS", "WWCSTAGE", "CONNECT", "Connect", "Connect");
            connectMergeRepl.RunDataSync();
            areAllInSync += 1; 
        }
        catch (Exception)
        {
            if (worker != null) worker.CancelAsync();
        }
    }

    static void matrixBW_DoWork(object sender, DoWorkEventArgs e)
    {
        BackgroundWorker worker = sender as BackgroundWorker;
        try
        {
            MergeRepl matrixMergeRepl = new MergeRepl(SystemInformation.ComputerName + "\\SQLEXPRESS", "WWCSTAGE", "MATRIX", "MATRIX", "MATRIX");
            matrixMergeRepl.RunDataSync();
            areAllInSync += 1; 
        }
        catch (Exception)
        {
            if (worker != null) worker.CancelAsync();
        }
    }

    static void aspBW_DoWork(object sender, DoWorkEventArgs e)
    {
        BackgroundWorker worker = sender as BackgroundWorker;
        try
        {
            MergeRepl aspnetdbMergeRepl = new MergeRepl(SystemInformation.ComputerName + "\\SQLEXPRESS", "WWCSTAGE", "aspnetdb", "aspnetdb", "aspnetdb");
            aspnetdbMergeRepl.RunDataSync();
            areAllInSync += 1; 
        }
        catch (Exception)
        {
            if (worker != null) worker.CancelAsync();
        }

    }

Answer 1

我只使用一个。

我认为BackgroundWorker允许我执行长时间运行的任务并保持UI响应。

如果我想要多个线程，我使用ThreadPool。

Answer 2

嗯，首先，我很抱歉这样说，但你的代码伤害了我的眼睛......

整个混乱可以像这样重写：

        internal static void SyncNow()
        {
            CreateWorker(new MergeRepl(SystemInformation.ComputerName + "\\SQLEXPRESS", "WWCSTAGE", "aspnetdb", "aspnetdb", "aspnetdb"));
            //etc...
        }

        private static void CreateWorker(MergeRepl repl)
        {
            BackgroundWorker connect = new BackgroundWorker { WorkerReportsProgress = false, WorkerSupportsCancellation = true };
            connect.DoWork += new DoWorkEventHandler(DoWork);

            if (connect.IsBusy != true)
                connect.RunWorkerAsync(repl);
        }

        private static void DoWork(object sender, DoWorkEventArgs e) 
        { 
            BackgroundWorker worker = sender as BackgroundWorker; 
            try 
            { 
                MergeRepl aspnetdbMergeRepl = e.Argument as MergeRepl;
                aspnetdbMergeRepl.RunDataSync(); 
                areAllInSync += 1; 
            } 
            catch (Exception) 
            { 
                if (worker != null) worker.CancelAsync(); 
            } 
        }

接下来，我会将ThreadPool用于这样的事情，这将确保只产生特定数量的线程来完成这类工作。

Answer 3

你需要更具体地说明'flakiness'是什么，它是如何表现出来的。如果我理解正确，那么在没有代理支持的情况下，您正试图手动驱动与RMO类的合并复制。

需要注意的一点是，SQL Express仅支持一个调度程序，因此向其添加多个worker（挂起请求）不会产生太大影响，它们只会堆积在可运行队列中并争夺一个CPU运行它们。

其次，我不确定RMO复制类（我假设你使用它）支持在多个并行实例中进行同步，所以我可能没有必要做更多的事情，每个db只有一个BackgroundWorker（我可能是这个错了，我不是任何一个RMO专家。

背景工人是否有限制？技术或常识

3 个答案: