Reocurring resource problems on Win6

C

carpe-diem

Perch
Reoccuring resource problems on Win6

I think this question needs a debate as I've filed a ticket before and it happens again.
-----
Helo, one of sites we host on Win6 (Drivers Planet) is having problems at least once a day for some weeks now.

It's either a OutOfMemoryException or this:

System resource exceeded.
Description: An unhandled exception occurred during the execution of the current web request. Please review the stack trace for more information about the error and where it originated in the code.

Exception Details: System.Data.OleDb.OleDbException: System resource exceeded.

Source Error:


Line 113: var rs = new DP.myRS(db);
Line 114:
Line 115: rs.Query("SELECT TOP 2 NewsID, NewsLink, NewsDate FROM News ORDER BY NewsDate DESC;")
Line 116: while (rs.reader.Read()) {
Line 117: news += (" <p style='padding:3px 2px 3px 2px;'>" + dateCompat(rs.reader.Item(2)))

Stack Trace:


[OleDbException (0x80004005): System resource exceeded.]
System.Data.OleDb.OleDbCommand.ExecuteCommandTextErrorHandling(Int32 hr) +41
System.Data.OleDb.OleDbCommand.ExecuteCommandTextForSingleResult(tagDBPARAMS dbParams, Object& executeResult) +174
System.Data.OleDb.OleDbCommand.ExecuteCommandText(Object& executeResult) +92
System.Data.OleDb.OleDbCommand.ExecuteCommand(CommandBehavior behavior, Object& executeResult) +65
System.Data.OleDb.OleDbCommand.ExecuteReaderInternal(CommandBehavior behavior, String method) +112
System.Data.OleDb.OleDbCommand.ExecuteReader(CommandBehavior behavior) +69
System.Data.OleDb.OleDbCommand.ExecuteReader() +7
DP.myRS.Query(String sql) +20
invoker0.Invoke(Object , Object[] ) +29
Microsoft.JScript.JSMethodInfo.Invoke(Object obj, BindingFlags options, Binder binder, Object[] parameters, CultureInfo culture) +126
Microsoft.JScript.LateBinding.CallOneOfTheMembers(MemberInfo[] members, Object[] arguments, Boolean construct, Object thisob, Binder binder, CultureInfo culture, String[] namedParameters, VsaEngine engine) +1080
Microsoft.JScript.LateBinding.Call(Binder binder, Object[] arguments, ParameterModifier[] modifiers, CultureInfo culture, String[] namedParameters, Boolean construct, Boolean brackets, VsaEngine engine) +2181
Microsoft.JScript.LateBinding.Call(Object[] arguments, Boolean construct, Boolean brackets, VsaEngine engine) +224
ASP.default_aspx.__Render__control1(HtmlTextWriter __output, Control parameterContainer) in D:\hshome\d-planet\driversplanet.com\default.aspx:115
System.Web.UI.Control.RenderChildren(HtmlTextWriter writer) +27
System.Web.UI.Control.Render(HtmlTextWriter writer) +7
System.Web.UI.Control.RenderControl(HtmlTextWriter writer) +243
System.Web.UI.Page.ProcessRequestMain() +1929
 
carpe-diem,

Firstly, I am very sorry for the reoccurance of this problems. I'd like to reassure you that we are on top of it. Stephen as well as myself have been working on this issue. We have isolated a few causes and are watching Win6 very closely. We will also be adding more RAM onto this machine within 24 to 48 hours.

EDIT: This issue is only affecting ASP.NET on Win6 and is occuring once in every 2 days.
 
Thank you for your support. It might be of interest to you that some other .NET sites we host elswhere have had problems similliar to this because the provider was running IIS6 in IIS5 compatibility mode which disabled several .NET optimisation functions and incurred overhead on .NET applications.
 
Win6 is a Windows 2000 IIS5 machine

We have discontinued using Windows 2003/IIS6 in the temporarilu due to the number of stability issues we faced with it when we introduced it this winter.
 
The problem reoccured today again on Win6 at Date/Time: Thu Jul 1 18:48:36 EDT 2004

The problem lasted for 15 minutes before we rebooted IIS. I deleted the two threads that were reporting the same issue. We are working very hard on this
 
Ok I believe this is resolved now. We have edited the machine.config to recycle if more than 45% of availabed memory is consumed. For the technical that would be this line in the machine.config:
memoryLimit="45"
 
OK, what appears to be happening is that there is a 1GB treshhold for the asp.net worker process. Once it uses that much of ram, it starts returning out of resource errors regardless of how much ram is on the system

In any case, hitting 1GB is highly abnormal considering our other servers are hovering at around 400 to 500MB. This indiciates that some customer's intensive asp.net application has a bad memory leak.

What we have done (as stephen explained) is configured the worker process to automatically recycle when it hits about .9GB. THat would prevent this error from ever occuring till we can locate the offending .net application

You can view asp.net's stats here:
http://win6.jodoshared.com/aspxstats.aspx
 
Thanks for the update. I will test my application with homer just to know if it could be my bad coding practices. Normal home operation is eating max 50 mb of memory, even if I click and refresh as mad. 8)

Update: After testing for an hour with 40 concurrent connections max memory used by aspnet_wp was 120 mb, then decreased.
 
You must log in or register to reply here.
Back
Top