We've had a WebCenter Spaces sites in production (Internet and Intranet) for about 2 weeks now. We have two physical servers clustered for Spaces and two for UCM. One thing that I've noticed is that in EM it shows that the active sessions all throughout the night stay at around 300,000. This doesn't really make any sense since there really shouldn't be more than *maybe* 100 people on during that time. The session time-out on both Spaces is set to 20 minutes. The servers can handle this load, but I don't understand why they should have to. Not to mention whenever we have Network issues and one of the spaces servers drop out for a few minutes all those active sessions are moved to the second server.. And 550,000+ "active sessions" seems a good bit high.
Active sessions, from what I understand, just means open connections to Spaces, not necessarily individual users. Is that correct? What can cause these sessions to stay as "active" even though everything times-out after 20 minutes? My theory is that someone in the site is a piece of code that isn't releasing the sessions that it uses, but before the developers have to go back through it all I'd like to exhaust all other options.
Please, see this article by Yannick Ongena and try to apply the recommendations posted there: http://www.ateam-oracle.com/high-availability-considerations-for-webcenter-portal-and-adf/
We've actually done all of the steps in that article. We went through load testing of 3000 users over 8 hours and didn't have a single issue. Then suddenly when we went live we were hit with these extremely high active sessions. This morning we were at 1.3Million active sessions on each spaces server!!!
So now I have two more questions:
1) Is there a way to kill sessions without restarting the managed server?
2) Does anyone have any suggestions on how to troubleshoot what is holding the session??