|
|
| Author |
Messages |
|
DNN Professor Gold Membership
 Ventrian Master Posts:764


 |
| 11/05/2007 9:05 PM |
|
Hi Scott,
Curious, have you had anybody report a memory leak with any of your recent module upgrades?
I have been battling a issue which WH4L has identified on my justnorth.com site. It is consuming memory like crazy.
I don't want to make this sound like it is any of your modules causing the issue. I am trying to isolate the problem. The reason I ask the question is that the issue began two weeks ago and the only modules I have added, in that period, are upgrades to your mods. Just trying to troubleshoot.
WH4L says:
Actually, I checked your server memory allocation, it seems like aspnet pool2 is using large memory, about 180mb and increasing, but it only serve 2 sites, jivemg.com and justnorth.com, I have seperate them into different pools and justnorth.com is actually the application that consume the most resource.
It is not really about sharing application pool problem though. I can create another application pool and seperate those 2 sites so you can identify them.
All the evidence pointing to justnorth.com, right after iisreset, justnorth used about 100mb and increasing fast as I browse / refresh the page.
Any ideas on how to troubleshoot this?
Best regards,
Buck |
|
Buck Anderson - DNN Professor
http://DNNprofessor.com
http://twitter.com/DNNprofessor
DotNetNuke Training - XMod Training Coach and XMod MVP
 |
|
|
Scott McCulloch Administrators
 Ventrian Master Posts:17204


 |
| 11/05/2007 9:23 PM |
|
Does it constantly increase exponentially and does not stop? Some of my app pools run at about 400mb during peak load. |
|
Scott McCulloch Site Administrator |
|
|
DNN Professor Gold Membership
 Ventrian Master Posts:764


 |
| 11/06/2007 6:33 PM |
|
Posted By smcculloch on 11/05/2007 9:23 PM
Does it constantly increase exponentially and does not stop? Some of my app pools run at about 400mb during peak load.
Good question. I will ask them.
The site has been down again all afternoon. They want me to purchase another 100 mb of memory to put a band aid on it until they can get the site stable.
It seems that all you have to do is make a dozen callbacks and the site goes down. |
|
Buck Anderson - DNN Professor
http://DNNprofessor.com
http://twitter.com/DNNprofessor
DotNetNuke Training - XMod Training Coach and XMod MVP
 |
|
|
Gary Registered Users
 Ventrian Addict Posts:66

 |
| 11/08/2007 7:48 AM |
|
In DNN 4.7.0, i know there were 2 memory leak issues that were fixed.
DNN-6000 Ajax.Isinstalled causes System.OutOfMemory exception
and
DNN-6372 CacheDependency Memory Leak
Maybe this will help you troubleshoot or even solve your problem. |
|
Gary |
|
|
Déclic Vidéo Registered Users
 Ventrian Master Posts:1283


 |
|
Simon Holman Gold Membership
Posts:10

 |
| 11/30/2007 12:58 PM |
|
Buck
We had a customer with a similar issue recently on one of our servers.
It turned out to be the GX-Chat and GX-WhosThere modules. Are you running either of those?
Your host should have tools to be able to tell which scripts are causing problems and chewing up memory.
Regards
Simon |
|
Expeed ASP.NET & MSSQL Hosting DotNetNuke Hosting
|
|
|
DNN Professor Gold Membership
 Ventrian Master Posts:764


 |
| 12/01/2007 1:05 PM |
|
Hi Simon,
Thanks for sharing the info.
Although I do not use either of the modules you mention, I am wondering if Scott's Who's On Line module might be having the same effect.
Any thoughts, Scott? |
|
Buck Anderson - DNN Professor
http://DNNprofessor.com
http://twitter.com/DNNprofessor
DotNetNuke Training - XMod Training Coach and XMod MVP
 |
|
|
DNN Professor Gold Membership
 Ventrian Master Posts:764


 |
| 12/01/2007 1:16 PM |
|
Posted By GHunter on 11/08/2007 7:48 AM
In DNN 4.7.0, i know there were 2 memory leak issues that were fixed.
DNN-6000 Ajax.Isinstalled causes System.OutOfMemory exception
and
DNN-6372 CacheDependency Memory Leak
Maybe this will help you troubleshoot or even solve your problem.
Hi Gary,
Thank you for the info. I have upgraded our two most trafficked sites to DNN 4.70 and added a 100mb of memory to each. This has seemed to curb the out-of-memory errors.
Before i upgraded to DNN 4.70 (last night), I was continually getting the following SQL database errors when performing administrative tasks. Most often when performing callbacks. I also see, from the log file, that is has been occurring when being spidered by Yahoo Slurp. Any information on this?
System.Exception: Unhandled Error: ---> System.Data.SqlClient.SqlException: A transport-level error has occurred when sending the request to the server. (provider: TCP Provider, error: 0 - An existing connection was forcibly closed by the remote host.) at System.Data.SqlClient.SqlConnection.OnError(SqlException exception, Boolean breakConnection) at System.Data.SqlClient.SqlInternalConnection.OnError(SqlException exception, Boolean breakConnection) at System.Data.SqlClient.TdsParser.ThrowExceptionAndWarning(TdsParserStateObject stateObj) at System.Data.SqlClient.TdsParserStateObject.ThrowExceptionAndWarning() at System.Data.SqlClient.TdsParserStateObject.WriteSni() at System.Data.SqlClient.TdsParserStateObject.WritePacket(Byte flushMode) at System.Data.SqlClient.TdsParserStateObject.ExecuteFlush() at System.Data.SqlClient.TdsParser.TdsExecuteRPC(_SqlRPC[] rpcArray, Int32 timeout, Boolean inSchema, SqlNotificationRequest notificationRequest, TdsParserStateObject stateObj) at System.Data.SqlClient.SqlCommand.RunExecuteReaderTds(CommandBehavior cmdBehavior, RunBehavior runBehavior, Boolean returnStream, Boolean async) at System.Data.SqlClient.SqlCommand.RunExecuteReader(CommandBehavior cmdBehavior, RunBehavior runBehavior, Boolean returnStream, String method, DbAsyncResult result) at System.Data.SqlClient.SqlCommand.RunExecuteReader(CommandBehavior cmdBehavior, RunBehavior runBehavior, Boolean returnStream, String method) at System.Data.SqlClient.SqlCommand.ExecuteReader(CommandBehavior behavior, String method) at System.Data.SqlClient.SqlCommand.ExecuteReader(CommandBehavior behavior) at Microsoft.ApplicationBlocks.Data.SqlHelper.ExecuteReader(SqlConnection connection, SqlTransaction transaction, CommandType commandType, String commandText, SqlParameter[] commandParameters, SqlConnectionOwnership connectionOwnership) at Microsoft.ApplicationBlocks.Data.SqlHelper.ExecuteReader(String connectionString, CommandType commandType, String commandText, SqlParameter[] commandParameters) at Microsoft.ApplicationBlocks.Data.SqlHelper.ExecuteReader(String connectionString, String spName, Object[] parameterValues) at DotNetNuke.Data.SqlDataProvider.GetTabModules(Int32 TabId) at DotNetNuke.Entities.Modules.ModuleController.GetTabModules(Int32 TabId) at DotNetNuke.Entities.Portals.PortalSettings.GetPortalSettings(Int32 TabId, PortalAliasInfo objPortalAliasInfo) at DotNetNuke.Entities.Portals.PortalSettings..ctor(Int32 tabId, PortalAliasInfo objPortalAliasInfo) at Ventrian.FriendlyUrl.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 ---
Best regards,
Buck |
|
Buck Anderson - DNN Professor
http://DNNprofessor.com
http://twitter.com/DNNprofessor
DotNetNuke Training - XMod Training Coach and XMod MVP
 |
|
|
Gary Registered Users
 Ventrian Addict Posts:66

 |
| 12/02/2007 8:34 AM |
|
Hi Buck,
That error is a General Error with TCP connections to the SQL Server and is very hard to troubleshoot. Its cause can be anything from networking issues to SQL Server Configuration to the way DNN is opening and closing connections to the SQL Server.
I am not sure how yahoo slurp spider would cause you to get that error. From what i have seen in my own logs is that it only gets spidered by a couple of spiders at one time, and shouldn't be hammering your site. Googlebot i have heard can get out of hand, hitting a site with thousands of spiders and can slow your site down or even bring it down.
So, is your site working ok now? Are you still getting that error after upgrading to DNN 4.7.0? |
|
Gary |
|
|
DNN Professor Gold Membership
 Ventrian Master Posts:764


 |
| 12/02/2007 5:19 PM |
|
Posted By GHunter on 12/02/2007 8:34 AM
Hi Buck,
That error is a General Error with TCP connections to the SQL Server and is very hard to troubleshoot. Its cause can be anything from networking issues to SQL Server Configuration to the way DNN is opening and closing connections to the SQL Server.
I am not sure how yahoo slurp spider would cause you to get that error. From what i have seen in my own logs is that it only gets spidered by a couple of spiders at one time, and shouldn't be hammering your site. Googlebot i have heard can get out of hand, hitting a site with thousands of spiders and can slow your site down or even bring it down.
So, is your site working ok now? Are you still getting that error after upgrading to DNN 4.7.0?
Hi Gary,
Again, thanks for taking the time to respond. I will be tracking this carefully. To date, the two sites that have been upgraded to DNN 4.7.0 have not prompted the error. |
|
Buck Anderson - DNN Professor
http://DNNprofessor.com
http://twitter.com/DNNprofessor
DotNetNuke Training - XMod Training Coach and XMod MVP
 |
|
|
Scott McCulloch Administrators
 Ventrian Master Posts:17204


 |
| 12/02/2007 5:28 PM |
|
Where are they uses the standard executeReader like most other DNN Modules:- Return CType(SqlHelper.ExecuteReader(ConnectionString, DatabaseOwner & ObjectQualifier & "DnnForge_WhereAreThey_GetOnlineUserLocations", portalId, GetNull(roles), hideSuperUsers), IDataReader) It doesn't use any ajax at all. Here is the one from the html/text module:- Return CType(SqlHelper.ExecuteReader(ConnectionString, DatabaseOwner & ObjectQualifier & "GetHtmlText", moduleId), IDataReader) |
|
Scott McCulloch Site Administrator |
|
|
DNN Professor Gold Membership
 Ventrian Master Posts:764


 |
| 12/02/2007 5:36 PM |
|
Posted By smcculloch on 12/02/2007 5:28 PM
Where are they uses the standard executeReader like most other DNN Modules:-
Return CType(SqlHelper.ExecuteReader(ConnectionString, DatabaseOwner & ObjectQualifier & "DnnForge_WhereAreThey_GetOnlineUserLocations", portalId, GetNull(roles), hideSuperUsers), IDataReader)
It doesn't use any ajax at all.
Here is the one from the html/text module:-
Return CType(SqlHelper.ExecuteReader(ConnectionString, DatabaseOwner & ObjectQualifier & "GetHtmlText", moduleId), IDataReader)
Hi Scott,
I am understanding your comment to say that 'Where Are They' can be ruled out as a cause of the problem, correct? |
|
Buck Anderson - DNN Professor
http://DNNprofessor.com
http://twitter.com/DNNprofessor
DotNetNuke Training - XMod Training Coach and XMod MVP
 |
|
|
Scott McCulloch Administrators
 Ventrian Master Posts:17204


 |
| 12/02/2007 6:03 PM |
|
| Yep, I think this can be ruled out, it's 2 stored procedure call using the standard calls. |
|
Scott McCulloch Site Administrator |
|
|
Déclic Vidéo Registered Users
 Ventrian Master Posts:1283


 |
| 12/06/2007 6:13 AM |
|
I join the club... Yesterday, WH4L informed me that my installation has a high consumption too... {...} I found that the issue should be caused by the application pool had used up all the resources for the site and can't recycle itself that cause the issue {...} it seems that the issue should be from your application, would you please give it a check on your application coding to see if it works properly without any issue? {...} If you find something Buck, keep me informed.... The old problems re-appeared... :-( |
|
Pinnacle Studio, proDAD and Boris FX tips and tricks, tutorials... Déclic Vidéo FX |
|
|
ROBAX Gold Membership
 Ventrian Master Posts:1018


 |
| 12/06/2007 3:39 PM |
|
One tip.. before I got remote desktop, I spent two months observing my account with Inventua's System Monoitoring kit: http://www.inventua.com/dnn-systemmonitor-overview.content It's free to try and I ended up buying it. (Just one comment about the developer: His habit of putting permanent advertising in the module is incredibly unprofessional, however he can't see that - go figure) In any case, the module lets you watch memory and cpu usage from the resources allocated to your account - put the main module on a page and refresh it now and then while you're working - or auto-refresh it. I was trying to see what would happen before something went wrong with the DNN portal, and it let me detect a few issues. Even so, applying the tweaks needed just wasn't working by going through the support desk, although I did try. Remote desktop has solved that... but what a drama and cost to go through to get DNN working smoothly, when all the way from a cheap shared account to the pricey VPS, my PHP carts have sat alongside DNN and not missed a beat. That issue where the site can't recycle at all seems to happen when it simply runs out of allocated memory. It freezes up and won't return until it has crashed itself. The only way to deal with it is tweaking the recycling triggers. If you don't let your portal reach the limit, it basically won't run into trouble... even with a bad module. You can see the memory exceptions in the application log in the module mentioned and also often in the DNN event viewer. I used to wonder - and asked the support desk more than once - why it doesn't just spill out into disk cache, but I never got an answer for that.. perhaps IIS doesn't use disk cache like that? Rob |
|
|
|
|
Déclic Vidéo Registered Users
 Ventrian Master Posts:1283


 |
| 12/07/2007 1:56 AM |
|
Wow robax, thanks a lot, looks great this inventua module.
One question, you said:
That issue where the site can't recycle at all seems to happen when it simply runs out of allocated memory. It freezes up and won't return until it has crashed itself. The only way to deal with it is tweaking the recycling triggers. If you don't let your portal reach the limit, it basically won't run into trouble... even with a bad module. You can see the memory exceptions in the application log in the module mentioned and also often in the DNN event viewer.
I am not sure that I have understood everything (I am french...). But it seems to me that it is exactly what's happenning. It can not recycle at all, running out of allocated memory and it freezes and do not return until it crash (sometimes it takes a huge amount of time, more than 30 minutes). So, HOW DO YOU TWEK the recycling trigger ????
DV |
|
Pinnacle Studio, proDAD and Boris FX tips and tricks, tutorials... Déclic Vidéo FX |
|
|
ROBAX Gold Membership
 Ventrian Master Posts:1018


 |
| 12/07/2007 2:53 AM |
|
Yes, what you're experiencing and even what Buck is having sounds a lot like what was happening to me. So long as you're on a shared account you won't be able to do anything about it. The best you can do is list all your domain names that are in the same portal, in a support ticket and tell them to make sure that they are all in the same application pool. If even one is in another pool, it will cause a significant increase in resource requirements. I was able to tell using that module that they had not placed one in the pool. If you're on a VPS, then you can specify exactly what you want to happen. If you look up IIS recycling options in IS help you'll see various trigger options and these can be set for each pool and some things even for each domain in the pool. The problem is that it needs watching and tweaking over time to see exactly what happens. I tried getting the support desk to set things for me and it didn't work. Once I had remote desktop I discovered that there are really so many ways of setting it up that there's no way it would ever work through a support monkey. I was mainly a bit miffed (not happy) that I had to get the VPS and remote far sooner than I actually wanted to, but it has been good. If you can get them to put your primary portal into its own pool with all domain in it, then tell them to set it up to recycle once every 3 hours. Don't ask for memory limit recycling as I found that it wasn't accurate and even got me into trouble when it got into a loop and was shutting down every 2-3 minutes. I'm not sure whether they'll let you have your own pools in a shared environment, but if they will then that will give you a good starting point. oh and i also stopped using keep-alives as it just doesn't make sense to waste resources on sites that aren't being visited. The popular sites stay alive fine and I don't care if the first visit to one of the others has to wait 15 seconds or so. Let me know what the system monitor module reports for memory and cpu type and whcih account you have. I'm interested. Rob |
|
|
|
|
Déclic Vidéo Registered Users
 Ventrian Master Posts:1283


 |
| 12/12/2007 5:47 AM |
|
Very interesting, thanks a lot !!
I have just sent you an email robax, I can not continue like that with WH4L, 1 hour downtime each day, not possible anymore.... I am very very sad, I thought they were very serious, no news for 4 days now... (2 tickets opened, many chat discussion with technician, seems they can not host DNN ..)
DV |
|
Pinnacle Studio, proDAD and Boris FX tips and tricks, tutorials... Déclic Vidéo FX |
|
|
ROBAX Gold Membership
 Ventrian Master Posts:1018


 |
| 12/12/2007 3:06 PM |
|
Sometimes they may simply not be able to help. What is the actual problem.. if the site simply spins away being extremely slow to the point where it hardly even works, it is probably running fine, but can't get going. That's basically what the problem was when I had it... the problem was lack of resources and apart form shifting me around different servers 3 times, there was nothing else they could offer with that account. I also ran several PHP carts and an Access cart plus a bunch of static sites on the same account and they all ran flawlessly and still do. DNN is simply a massive resource pig and webshosts can't simply fit a bunch of 600mb RAM DNN portals into the same cheapie accounts that were previously only hosting 30mb RAM PHP sites. I expect it has caught a lot of hosts off-guard. Places like PowerDNN restrict you to one DNN instance and one db for a reason. WH4L lets you have as many as you like.. but beware. If I had only one single site to run and it was not trying to grow it then I would have moved to a host who does shared DNN hosting better eg WSS or PowerDNN, but with my plans and much broader requirements, I decided to stay. I'll wait till I hear back in email from you, but I think that video bandwidth requirements may be a bigger concern for you if you're planning on hosting clips and growing your audience. You might find that someone entirely different is needed or you might have to plan for the fully dedicated. Rob |
|
|
|
|
Déclic Vidéo Registered Users
 Ventrian Master Posts:1283


 |
|
|
| You are not authorized to post a reply. |
|
|
|
ActiveForums 3.7
|
|