> We are running a loadtesting for CAS with 4 JBOSS servers. When user load is > increased to 400 we are getting an exception as below.
The exception you mentioned has been documented pretty thoroughly in http://www.ja-sig.org/issues/browse/CAS-782. Note that we have observed that deadlocks dramatically decrease if not disappear when the indices on the foreign keys are created as described in the issue. > The quartz job is running from all locations. > What is the best option to run quartz job in this scenario. > Only enabling Quartz job in one of the servers only? There are two recommended solutions at present: 1. Run the Quartz job on only one node. 2. Handle ticket cleanup on the database side. I would argue both solutions have significant downsides. The first option is bad for a high availability environment. The second is bad since it's not possible to determine the exact ticket state from the data in the ticket tables. I have developed an alternative ticket cleaner implementation that I am planning to attach to a Jira issue in the next week that allows the Quartz job to be deployed to all nodes but only one node at a time can execute. This is the ideal solution in my opinion. We expect deadlocks to disappear entirely once this goes into production. M -- You are currently subscribed to cas-dev@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-dev
