I'm back reading the document at https://wiki.jasig.org/display/CASUM/Clustering+CAS which creates some questions.
Assume I'm running 3.4.5 (or 3.4.6 - question is which is preferred at this point for my situation of multiple servers behind a LB, with everything stored in a shared database)? Do I need to go through all that is listed for Clustering to run CAS on multiple servers, or can I simply rely on having the data stored in a DB shared between the servers to have CAS work properly? Are sticky-sessions a requirement? What if a service hits a different CAS server than the one a user get's their ticket from? I know these are basic questions, and, feel free to point me at some references, or ignore me. thanks, chris On Apr 21, 2011, at 11:48 AM, Marvin Addison wrote: >> Note that CAS 3.4.7 fails behind a load balancer if you're not using sticky >> sessions but using clustered sessions (its because each server has its own >> key). >> Marvin's put in a fix to not have to use keys and I'll be cutting the >> release later. > > It's important to note that my fix would still require session > affinity since the random component of the flow key is stored in > Webflow conversation state. The requirement for session affinity, > however, is nothing new. When SWF finally re-introduces sessionless > flows, we can document how to do configure CAS for a non-sticky setup. > Until then, sticky is a requirement for clustered CAS. > > M > > -- > You are currently subscribed to [email protected] as: [email protected] > To unsubscribe, change settings or access archives, see > http://www.ja-sig.org/wiki/display/JSG/cas-user -- You are currently subscribed to [email protected] as: [email protected] To unsubscribe, change settings or access archives, see http://www.ja-sig.org/wiki/display/JSG/cas-user
