Bruce,
Unfortunately I fixed the issue by changing to a backup of the database - but I've run into a new issue - when the site is first hit - we see a HUGE memory spike upwards of 900 Megs by the w3wp.
When the provider is first building the cache is it building the cache for ALL the portals in an instance for just the one that's being requested? In our system we have 200 + portals and 76774 tabs. Each individual portal doesn't see a large amount of traffic on a daily basis. The problem we are having is it's takeing FOREVER to get the first site loaded after an application restart. On the order of 30 minutes MINIMUM.
We ran into this problem with the stock DNN core code, turns out there's a section where the portal ID isn't being passed down to a lower level call when asking for tabs, so it would return ALL the tabs for ALL the portals with every call.. Not a problem once or twice, but do that a bunch of times and performance TANKS.
Is it possible to get the source code to review what's happening?
Thanks.