Hi team,

Testing a webfarm using DNN 4.4.1 - two web heads running IIS6 talking to a third SQL 2005 Std / Content server.

Sites on the farm work ok albeit very slowly (up to 30 sec for first page) and when load/stress testing the following exception occurs repeatedly in the DNN event viewer

--------------------

AssemblyVersion: 04.04.01
PortalID: -1
PortalName:
UserID: -1
UserName:
ActiveTabID: -1
ActiveTabName:
RawURL: /aboutthepeople/tabid/55/Default.aspx
AbsoluteURL: /Default.aspx
AbsoluteURLReferrer: http://www.xxxxxx/aboutwhatwedo/tabid/53/Default.aspx
UserAgent: Mozilla/4.0 (compatible; MSIE 6.0; Windows NT 5.1; SV1; .NET CLR 1.1.4322; .NET CLR 2.0.50727; InfoPath.1)
DefaultDataProvider: DotNetNuke.Data.SqlDataProvider, DotNetNuke.SqlDataProvider
ExceptionGUID: 90a3c304-a8b1-494b-9c77-fd87ad28fbe3
InnerException: Unhandled Error:
FileName:
FileLineNumber: 0
FileColumnNumber: 0
Method: System.IO.__Error.WinIOError
StackTrace:
Message: System.Exception: Unhandled Error: ---> System.IO.IOException: The process cannot access the file '\\xxxsql1\wwwroot\DNN441_xxxx\Portals\_default\Cache\VGFiczA=.resources' because it is being used by another process. at System.IO.__Error.WinIOError(Int32 errorCode, String maybeFullPath) at System.IO.File.Delete(String path) at DotNetNuke.Services.Cache.FileBasedCachingProvider.FBCachingProvider.DeleteCacheFile(String FileName) at DotNetNuke.Services.Cache.FileBasedCachingProvider.FBCachingProvider.GetPersistentCacheItem(String CacheKey, Type objType) at DotNetNuke.Entities.Tabs.TabController.GetTabsByPortal(Int32 PortalId) at DotNetNuke.Entities.Tabs.TabController.GetTab(Int32 TabId, Int32 PortalId, Boolean ignoreCache) at DotNetNuke.Entities.Portals.PortalSettings.GetPortalByTab(Int32 TabID, String PortalAlias) at DotNetNuke.HttpModules.UrlRewriteModule.OnBeginRequest(Object s, EventArgs e) at System.Web.HttpApplication.SyncEventExecutionStep.System.Web.HttpApplication.IExecutionStep.Execute() at System.Web.HttpApplication.ExecuteStep(IExecutionStep step, Boolean& completedSynchronously) --- End of inner exception stack trace ---
Source:
Server Name: xxxxx

This error appears to relate to file access contention by the two IIS servers accessing a common portals\default\cache folder which is now being used on account of 'enablewebfarmsupport=true'

Didn't have this problem when IIS+SQL+Content were on the same box nor does it happen in the web farm unless multiple requests are generated by the stress tester app' (MWAST)

Error sounds a bit ominous and haven't been able to reference on forum searches - any ideas on what could be wrong?

Also... is there any benefit / problem with turning off IIS ASP caching or changing to IIS ASP cache to memory... does IIS ASP caching interfere with the DNN FBCaching system?

Yes - web heads are using UNC path to the file server... transferred a couple of low volume sites to the web farm which run ok albeit v slowly...  need to get page load times down to < 5 sec and resolve the aforementioned FBCaching error

Wondered if the FBCaching error may be the result of using the Microsoft Web Stress testing tool i.e. could thousands of requests from a single source in a short period of time cause some sort of cache file access contention?

Checked out your (excellent) site which loads quickly even from over the ditch... do you use app's like PageBlaster or is GZip sufficient?

Have 12 other DNN sites sitting on several servers inc. one very high volume site which can't be moved until performance issues are resolved ...

Preliminary testing seems to show that disabling IIS ASP Caching and DNN Cache Persistence in web.config makes things run a bit quicker... red herring perhaps... will test further

Any performance boosting tips would be much appreciated.

Thanks again - enabled IIS HTTP compression as suggested i.e. ticked compress application files but left compress static files unticked...

Wondered if there was any benefit/problem in a web farm environment of enabling the 'Auto-Sync File System' under Host > Advanced Settings > Other Settings and whether selecting 'Memory' as the Module Caching method under Host > Performance Settings was ok for a webfarm

Tested the site using Mercury LoadRunner and it generated dozens of the aforementioned errors i.e.

System.Exception: Unhandled Error: ---> System.UnauthorizedAccessException: Access to the path '\\xxxsql\wwwroot\DNN4_xxx\Portals\_default\Cache\VGFiUGVybWlzc2lvbnMw.resources' is denied. at System.IO.__Error.WinIOError(Int32 errorCode, String maybeFullPath)

Noticed similar requests for help in a couple of other ASP/DNN forums where this error appears under high load conditions

Under 'light' load still getting about 7 of the aforementioned* error messages for every 1000 visitors

Another question - page 12 of the DNN Web Farm document seems to imply that the following schedule items should by run by only one web server in the farm

DotNetNuke.Entities.Users.PurgeUsersOnline
DotNetNuke.Services.Log.SiteLog.PurgeSiteLog
DotNetNuke.Services.Scheduling.PurgeScheduleHistory
DotNetNuke.Services.Log.EventLog.PurgeLogBuffer
DotNetNuke.Services.Log.EventLog.SendLogNotifications
DotNetNuke.Services.Search.SearchEngineScheduler
DotNetNuke.Modules.Admin.ResourceInstaller.InstallResources  
DotNetNuke.Services.FileSystem.SynchronizeFileSystem

With the exception of

DotNetNuke.Services.Cache.PurgeCache

which has the 'Run on Servers' box left empty so all web servers run it

Is this the correct configuration for Schedule items?