Every once in a while my DNN installations need to be kick-started.  I am not sure what is going wrong, but I have tried many things to get the stability.  Telltale sign is ususally that half of the site's main menu items are missing and logging in only brings up an error page. 

To get the site back up and running I just have to recycle my application pool in IIS.  If I am able to log in to a portal as Host, I can go to Host Settings and restart the application which does the same thing.

I read an article somewhere on setting up a Web Garden by increasing the Worker Processes in the Performance tab in the Application Pool for in IIS.  This seemed to pretty much eliminate the issue of the DNN sites ocasionally going partially offline so I was a happy guy.

I later noticed some other strange things.  I tried to create a sub-portal under my parent portal and when I did, the new portal adopted the skin of the parent portal and when I would log into the Child portal to administer, I was actually logging into the parent portal.  I could get around this problem by making the new portal the sub-portal of another domain other than the parent domain.

Another thing that was happening (and still might) is that when I did create the new sub-portal and log in as Host, and change the skin, the skin of the parent portal would also be changed.  If I logged in as admin, I never had this problem.

Not knowing what was causing all this eratic behaviour, I decided to upgrade from DNN 4.6.2 to 4.8.  I usually don't have issues when upgrading, but this time was hell.  At first I thought all was well until I got a call from a customer with a bilingual portal that was no longer bilingual.  My DSLocalizator was not working right.  So I found that I just had to install the "Hack" and the site was once again bilingual.

The next thing that I noticed was that the sites were nearly impossible to maintain.  I would make a settings change to a module, return to the page and the change would not be there.  So I would refresh and the changes would show, refresh again and they were gone. Then I would get an error message of Failed ViewState.  My event log was full of Scheduler errors,  It was really a lousy deal.  I though it was an issue with DSLocalizator and went to work getting it configured properly. 

I thought all my problems were behind me, everything seemed to work good, but the strange administration issues kept coming up.  I finally remembered the web-garden setting in IIS and put it all together.  I went back to 1 and now my strange problems are gone.

Now if only I had a way to recycle an Application when a site goes half way down.

Thanks for listening, had to get this out there as nobody seems to have ever seen anything like this.

Mike

the same thing is happening to my sites. any solutions?

I have no modules that I developed myself and about 30 commercial modules including ventrian modules.  Many of the modules have satellite modules like Simple Gallery and Ultra Media Gallery and if I counted all the satellite modules, there are more than 40.

I do seem to have caching problems still.  I recently updated a page for someone using Text/Html and the changes did not appear right away.  I thought that problem was gone.  I keep putting the blame on DSLocalizator, it has been a bear since upgrading to 4.8

Mike

Mike,
We just went through this with our 4.8 upgrade. The behavior was essentially no response from the server. IISRESET was the only way out. If you have the ability to watch performance, processor % will drop to zero when this happens. It turns out there is a bug in the DNN 4.8 scheduler which runs the scheduled tasks over and over and over. In our case we looked at schedule history and saw items like PurgeUsersOnline running 1000 times a second!

There is a documented bug in scheduler that causes this. It will be resolved in 4.8.1. To get around it, we turned off all scheduled jobs in DNN. Things are now running smoothly again.