On Sep 11, 2013, at 9:24 AM, "St Laurent, Mark" <mark.stlaur...@yc.edu> wrote:
> So, I have been trying to implement Ehcache on our CAS cluster, and it is
> mostly working. It caches tickets, and the two nodes are replicating and
> validating each other’s tickets. But it still has an issue with availability.
> Here is what is happening:
>
> 1. I open a session with a browser and authenticate with a CASified
> application,
> 2. I disconnect one of the CAS servers in the cluster to simulate a
> failure,
> 3. When I try to connect to another CASified application, it requires
> me to reauthenticate. If I look in the logs there is a single entry that says
> SERVICE_TICKET_NOT_CREATED.
>
> How do I get the cluster to continue validating service tickets after a node
> has failed?
Two possibilities come to mind offhand:
1. In the original Ehcache example TGTs were not being replicated due to a
missing bean directive. Ensure the TGT bean is inheriting from the abstract
ticket cache:
<bean id="ticketGrantingTicketsCache"
class="org.springframework.cache.ehcache.EhCacheFactoryBean"
parent="abstractTicketCache">
2. The TGT cache does not by default replicate synchronously. Depending on the
timing of your simulated failure, the client TGT may not have replicated to its
peer node(s).
Tom.
--
You are currently subscribed to cas-user@lists.jasig.org as:
arch...@mail-archive.com
To unsubscribe, change settings or access archives, see
http://www.ja-sig.org/wiki/display/JSG/cas-user
