Hi Juan, We have been running an architecture almost identical to what you're after except we're not using sticky sessions. Raw facts are: * its been production for circa 2 years * binaries are stored on an NFS mount shared by all servers using the JackRabbit DataStore feature * node data is stored in a central MySql database * JackRabbit clustering is enabled to keep the nodes in sync.
The above config has worked well for us. Regards, Shaun -----Original Message----- From: Juan Pereyra [mailto:[email protected]] Sent: 30 June 2009 18:10 To: [email protected] Subject: Jackrabbit clustering Hi guys, We're developing a system that uses jackrabbit as its backend and we hope to be able to run it in a clusterized environment, that is, basically using a load balancer with sticky session. Also, we were hoping to be able to use the database to store simple properties and the filesystem to store the actual binary content that our users upload. In the wiki (http://wiki.apache.org/jackrabbit/Clustering) it says: "The persistence manager needs to be transactional, and need to support concurrent access from multiple processes. When using Jackrabbit, one option is to use a database persistence manager, and use a database that does support concurrent access." So, what are the options to run in a clusterized environment? 1) Storing everything in the db, which would force us to use something like Oracle. 2) Store just the references to the files and handle the transactions ourselves? 3) Any other ideas? Also, it'd be great if someone else did face this problem before and what did you ended up doing. Thanks a lot! Juan
