Search results matching tag 'performance'

Random ASP.NET 2.0 crashes under load (A field or property with the name X was not found on the selected data source)

jimjimjimjim — Tue, 24 Nov 2009 05:00:00 GMT

We have an application that is an extension of Nav to the web. The system is used for accounting and approval of invoices/credit memos. The app is written in C#, and is run on an IIS webserver. The connection to Nav is via MSMQ and NAS. The web makes a request XML-style to Nav, and gets the result back as an xml message.

The system works fine until it gets under some load (50-100 users). When that happens we get errors in the event viewer like "A field or property with the name X was not found on the selected data source". This happens with different fields and on different pages, but it is always the same field on a particular page. I e F68 on PurchDoc.aspx, F3 on PurchDocLines.aspx, etc.

When I check the input/output xml messages from Nav theyre completly OK. Something happens on the website when it tries to read the message. I've posted this on Dynamics Nav forums aswell, but it is ASP .NET 2.0 that raises the error and the input data is ok to the asp .net application so I thought I´d post it here.

I use Visual Studio 2008. The pages are all built with ObjectDataSource controls. I receive the response from Nav with

System.Messaging.Message inMsg;
XmlDocument doc;
inMsg = mqFromNavision.Receive(new TimeSpan(0, 0, mqTimeOut)); 
doc = (XmlDocument)(inMsg.Body);

An example XML response is

- <Response xmlns:xs="http://www.w3.org/2001/XMLSchema" xmlns:msdata="urn:schemas-microsoft-com:xml-msdata">
- <xs:schema>
- <xs:element name="NavisionDataSet" msdata:IsDataSet="true">
- <xs:complexType>
- <xs:choice maxOccurs="unbounded">
- <xs:element name="T51000">
- <xs:complexType>
- <xs:sequence>
  <xs:element name="F1" type="xs:string" minOccurs="0" /> 
  <xs:element name="F2" type="xs:string" minOccurs="0" /> 
  <xs:element name="F5" type="xs:string" minOccurs="0" /> 
  </xs:sequence>
  </xs:complexType>
  </xs:element>
  </xs:choice>
  </xs:complexType>
  </xs:element>
  </xs:schema>
- <NavisionDataSet>
- <T51000>
  <F1>TESTUSER</F1> 
  <F2>Test</F2> 
  <F5>Närvarande</F5> 
  </T51000>
  </NavisionDataSet>
  <messageid>8589</messageid> 
  </Response>

Here, it would work most of the times when under load, and all of the times when not under load. But randomly it crashes under load and raises the error "A field or property with the name F2 was not found on the selected data source". Like I said earlier this happens on other pages too, with different XML data.

Have anyone experienced or heard of a similar issue? Could I try to run the code on .net 3.5 instead?

Reducing size of ASP.NET pages

ajay.ms2000 — Sat, 21 Nov 2009 05:00:00 GMT

We all know when a page is requested, ASP.NET process the page, its server controls and finally sends the Html to the client side for the browser to render it. The time taken to download the html to client side depends mainly on its final size. If your page is data rich, it would take much time to fetch it. So here I am presenting a technique to reduce the sizes of the pages.

When you see the pages html source by clicking the “View Source” in the browser, you could see that there are a lot of white spaces on the left side of each line on the html. This is actually a waste. Try saving the html page to your desktop and notice its size. Then delete all the spaces on the left side and then watch its size. You could see that the size of the page reduces considerably, sometimes more than 50%!! (watch the view source of Orkut in the browser).

Here is a technique to achieve the same.

1. Create a class in the App_Code, that is deriving from “System.Web.UI.Page” class.

public class MyPageBase : System.Web.UI.Page
{
}

2. Replace “System.Web.UI.Page” from all your web pages (.aspx.cs) and put MyPageBase. Means now all your webpages are deriving from MyPageBase.

3. Override the Render function in this class

public class MyPageBase : System.Web.UI.Page
{
    protected override void Render(HtmlTextWriter writer)
    {
    }
}

4. Place the below code inside the function and also write a support function as below

public class MyPageBase : System.Web.UI.Page
{
    protected override void Render(HtmlTextWriter writer)
    {
         using (HtmlTextWriter htmlwriter = new HtmlTextWriter(new StringWriter()))
        {

            base.Render(htmlwriter);
            string html = htmlwriter.InnerWriter.ToString().Trim();

            if (ConfigurationManager.AppSettings["OptimizeHtmlOutput"] != "0")
            {
                bool isAsync = !html.StartsWith("<");

                if (!isAsync)
                {
                    StringBuilder sb = MyPageBase._TrimHtml(html);
                    writer.Write(sb.ToString());
                }
                else
                {
                    StringBuilder sb = new StringBuilder();

                    int startIx = 0;
                    while (true)
                    {
                        int x = html.IndexOf("|updatePanel|", startIx);

                        if (x > -1)
                        {
                            int xS = html.LastIndexOf("\r\n", x); xS = (xS < 0 ? 0 : xS);
                            int xE = html.IndexOf("\r\n", x) + 2;

                            string header = html.Substring(xS, xE - xS);
                            header = header.Trim().TrimStart('|', ' ');

                            string sLen = header.Substring(0, header.IndexOf('|'));
                            int cLen = int.Parse(sLen);

                            string content = html.Substring(xE - 2, cLen);
                            content = MyPageBase._TrimHtml(content).ToString().Trim();

                            startIx = xE - 2 + cLen;

                            cLen = content.Length + 4;
                            header = (xS > 0 ? "|" : string.Empty) + cLen.ToString() + header.Substring(header.IndexOf('|'));

                            sb.AppendLine(header);
                            sb.AppendLine(content);
                        }
                        else
                        {
                            string stateData = html.Substring(startIx);
                            sb.Append(stateData.Trim());

                            writer.Write(sb.ToString());
                            break;
                        }
                    }
                }
            }
            else
            {
                writer.Write(html);
            }
        }
    }

}

    private static StringBuilder _TrimHtml(string source)
    {
        StringBuilder sb = new StringBuilder();

        source = source.Trim();
        using (StringReader sr = new StringReader(source))
        {
            string data = string.Empty;
            while (data != null)
            {
                data = sr.ReadLine();
                if (data != null)
                {
                    data = data.TrimStart(' ', '\t');
                    if (data != string.Empty) sb.AppendLine(data);
                }
            }
        }

        return sb;
    }
}

Notes:
1. Notice the usage of ConfigurationManager.AppSettings["OptimizeHtmlOutput"] != "0". You can define a key in app settings section of the web.config file to enable or disable this feature. Any value other that “0” will enable the optimization.
2. This code works well in AJAX based applications as well.

Compare the size difference and performance by changing the flag. This is very beneficial especially in pages which contains a huge amount of data.

Re: HOW TO USE BEST IDE FOR WEB DEVELOPMENT ON WINDOW PLATEFORM

Critters — Fri, 20 Nov 2009 05:00:00 GMT

1GB of RAM is pretty low these days, but if you can't afford more RAM, make sure to adjust your system for performance. Turn off any of the special effects in the system UI; if you have two drives, move your page file to a drive separate from the OS, and turn off Intel SpeedStep in your BIOS.

TransactionScope

vishalmarneni — Mon, 16 Nov 2009 05:00:00 GMT

I am trying to find out how using "Transaction Scope" in C# Code would effect performance.

Is there any tool / application I could use find out if using TransactionScope would reduce the performance of the application????

I have about 400users using the application.

Thanks,

Vishal

Which code snippet is better?

Ahmish — Mon, 02 Nov 2009 05:00:00 GMT

Which code snippet is better?

Code Snipped - 1

       DataTable dtLocal = (DataTable)this.ViewState["DataTable"];
        foreach (DataRow datarow in dtLocal.Rows)
        {
            //Do some operations
        }

Code Snippet -  2        
       foreach (DataRow datarow in ((DataTable)this.ViewState["DataTable"]).Rows)
        {
            //Do some operations
        }

Performance problems with HMC 4.5

dynamosympathisant86 — Wed, 07 Oct 2009 04:00:00 GMT

Hi,

we have some performance problems with a HMC 4.5 environment.

Every request against the MPS takes about 30seconds. I traced the requests and I realized that every operation against the Active Diretory takes a long time. Every other action is faster.

The problem is not based on hardware problems. It is an Windows Server 2003 environment and the DomainController has enough memory and CPU power.

Does everyone has an idea how to fix this performance problem? This problem is unacceptable at our frontend.

Best regards,

Ken

Rijndael performance issue

Vadim Milichev — Thu, 24 Sep 2009 04:00:00 GMT

We need to cipher users' files in our application.
The following is an encryption prototype method (in the real life it will accept an input stream). It is straightforward but we are not fully satisfied with the performance of this sample and considering changing the method used.

Is it possible to squeeze out some seconds from this code?

private static void EncryptFile(string fileName, string encryptedFileName, byte[] key, byte[] iv)
{
	using (RijndaelManaged rijndael = new RijndaelManaged {Key = key, IV = iv})
	{
		using (FileStream outFile = new FileStream(encryptedFileName, FileMode.CreateNew, FileAccess.Write, FileShare.None))
		{
			using (CryptoStream cryptoStream = new CryptoStream(outFile, rijndael.CreateEncryptor(), CryptoStreamMode.Write))
			{
				using (FileStream inFile = new FileStream(fileName, FileMode.Open, FileAccess.Read, FileShare.None))
				{
					byte[] buffer = new byte[bufferSize];
					int read;
					while ((read = inFile.Read(buffer, 0, bufferSize)) > 0)
					{
						cryptoStream.Write(buffer, 0, read);
					}
				}
			}
		}
	}
}

IF EXISTS() runs slower than straight Select *

Queuetip — Mon, 14 Sep 2009 04:00:00 GMT

On SQL SERVER 2000,

I've got a simple stored procedure that checks if a user has any new messages that they have not viewed. A straight "SELECT *" of all of their new messages runs quickly. However, when there are no new messages, an IF EXISTS(SELECT *...) takes a very long time. Much longer than if I just selected the data.

Here is a snippet of what the code looks like. I tried to make the table and column names as logically accurate as possible.

IF exists (SELECT *
		FROM Users A
			join UserGroupsLink B on B.UserID = A.ID
			join Groups C on C.ID = B.GroupID 
			join GroupMembers D on D.GroupID = C.ID
			join Messages E on E.MemberID = D.MemberID
			join MessageTypes F on F.ID = E.MessageTypeID
			LEFT join Messages_Viewed G ON G.MessageTypeID = F.ID
		WHERE
			A.ID = @UserID
			and F.IsActive = 'YES'
			and G.UserId is null)
	
		select 1 as 'HasNewMessages'
	ELSE
		select 0 as 'HasNewMessages'

The LEFT JOIN on that viewed table is used to grab all viewed messages, and only take the records that don't have a viewed message associated with it. I'm still not sure whether this is the best or most efficient logic. If anyone has an another solution to this scenario, I'd love to hear.

Anyway, my main question is, why is the IF EXISTS() statement taking far longer than the actual SELECT * statement that it is using? Should worst case scenario be the SELECT * returns nothing and it makes 1 extra step to finish the conditional statement? No where near what I'm seeing (148,206 reads for IF EXISTS() as opposed to 12,340 reads for SELECT *). The execution plans DO look different, but I'm still not sure why. Could it be JOIN ordering? Indexes? What makes EXISTS(SELECT *) different from straight SELECT *???? Thanks for any help.

IF exists (SELECT *
       FROM Users A
           join UserGroupsLink B on B.UserID = A.ID
           join Groups C on C.ID = B.GroupID
           join GroupMembers D on D.GroupID = C.ID
           join Messages E on E.MemberID = D.MemberID
           join MessageTypes F on F.ID = E.MessageTypeID
           LEFT join Messages_Viewed G ON G.MessageTypeID = F.ID
       WHERE
           A.ID = @UserID
           and F.IsActive = 'YES'
           and G.UserId is null)

       select 1 as 'HasNewMessages'
   ELSE
       select 0 as 'HasNewMessages'

Performance issue

vijaygarg — Thu, 10 Sep 2009 04:00:00 GMT

Env: Visual Studio 2008 (not sp1), IIS 5.1/6, Oracle 10G

We are having performance issue in our application and need to look into more datail where the issue is. Is there any profiling tool available from Microsoft or otherwise that can provide us details on where we should look into or which step/process is taking more time.

Thanks

Performance Problems

Marc81Nbg — Mon, 07 Sep 2009 04:00:00 GMT

Hi Guys,

i do have a performance problem while looping through a datasource. The DataSource comes from a DataSet and this one access a View from a SQL2005 Server (thought that would be faster than accessing the data with a huge sql inside the dataset - but is there a big difference?).

So know i have data inside my DataSet (1600Rows) which have to be loaded with some detailed data, the dataset could not provide. But exactly that looping and dataloading is to slow. Here is my code:

  For Each row In DataTable


            If prevID = row.IDAnfrage Then 'DoppelteIDs löschen
                row.Delete()
            Else
                prevID = row.IDAnfrage 'neue ID setzen für den nächsten Loop
            
                      '-----------------
        'Angebotsversion berechnen
        '-----------------
        Angebotsversion = 0
        StoredProcedure.uspGetAngebotsVersionFromAnfrageID(prevID, Angebotsversion)
        row.Angebotsversion = Angebotsversion


        '-----------------
        'Bewertungskennziffer errechnen
        '-----------------
        row.Bewertungskennziffer = GetPerformanceIndicatorByIDAnfrage(prevID)


        '-----------------
        'Ranking
        '-----------------
        ranking = 0
        Try : StoredProcedure.uspGetRankingFromAnfrageID(prevID, ranking)
        Catch ex As Exception
        End Try

        If ranking = 0 Then
            row.Ranking = row.Bewertungskennziffer
        Else
            row.Ranking = ranking
        End If




        '-----------------
        'Produktmanager
        '-----------------
        Produktmanager = ""
        Try : StoredProcedure.uspGetProductManagerFromAnfrageID(prevID, Produktmanager)
        Catch ex As Exception
        End Try

        If Produktmanager <> "" Then row.Produktmanager = Produktmanager


        '-----------------
        'Produktentwickler
        '-----------------
        Produktdeveloper = ""
        Try : StoredProcedure.uspGetProductDeveloperFromAnfrageID(prevID, Produktdeveloper)
        Catch ex As Exception
        End Try

        If Produktdeveloper <> "" Then row.Produktentwickler = Produktdeveloper


        '-----------------
        'Volumen
        '-----------------            
        Try : row.Volumen = GetInquiryVolumeByIDAnfrage(prevID)
        Catch ex As Exception
        End Try

        ''Timestamp
        'Try
        '    Dim Times As Date = row.Timestamp
        'Catch ex As Exception
        '    'ok - dbnull
        '    row.Timestamp = row.DatumAnfrage
        'End Try



            End If
        Next

So every single dataloading was sourced out to a stored procedure which have to be very fast! But - in my case this For Each Loop is to slow (about 20 Seconds for 1600 Rows).

Which looping would you prevere? Do i use a slow programming technique?
I would be happy if you could help me.

Thanks in advance

Marcus