We are experiencing errors with one of the CAS servers. If you receive an error similar to "Spring Framework used by CAS threw an unexpected exception", please refresh your page and try logging in again. The CAS admins are aware of the issue and are working to resolve the problem.
A core piece of networking equipment failed (one of the ServerIron load balancer nodes), and the transition to the failover load balancer caused connections to core services (Enterprise Directory, PostgreSQL database) to be interrupted. There are three CAS nodes and all but one recovered automatically; the third had to be restarted to restore all connectivity. This explains the transient nature of the problem; some folks were bound to the bad node and had problems while others never saw it.
I should note that we are planning a deployment in the near future to remove the database system dependency. That will be one fewer component to babysit when things inevitably go bad; we'll have a simpler and more reliable SSO service as an outcome.