kilburb wrote:ok I have a question. Looking through everything it seems that the only reason that we could not run TS in our clustered multi-server production systems is the fact that it uploads attachments to the files system. Now this could be over come by using a NFS mounted file system on all the servers that the files are uploaded to. Are there any other reasons that would prevent us from using TS in a clustered setup?
Do you plan to use several TrackStudio instances with single database ?
This configuration is not supported now because TrackStudio 2.8 does not support distributed caches (because JCS implementation does not work). For example, if one user move task on the first instance, the cache should be updated/invalidated on second instance too, but JCS can't work this way. JCS is only supported cache in Hibernate 2.0 and we can't solve this problem in 2.8.
But Hibernate 2.1 coming soon (beta 6 already available) - it supports distributed cache and can use not only JCS, but also OSCache, SwarmCache and commercial Coherence cache. After Gavin release 2.1 we'll switch to it and (I think) switch from JCS to SwarmCache. I hope that TrackStudio 3.0 should be cluster-ready.
PS. TrackStudio works fine under load even on single-node, it has been optimized for large amount of tasks/users/filters/workflows/etc.