Dear All, I'm having problems with load-balancing a single island of two orion appservers. I have followed the instructions from the clustering 'howto' document, and I've made changes to the relevant XML files. Note, I made the changes to the default web app files, rather than to the XML files specific to my app. From the documentation I expect the changes to therefore impact *all* apps, and that I should experience the desired clustering and load balancing for my app. Using the Microsoft Web App Stress Tool, running 100 threads which continually request a single page from the clustered app via the load balancer, things seem to work fine for the first minute or so, despite the fact that the balancer 'drops' servers periodically but then manages to re-discover them. The performance meter for the servers in the app cluster show that both machines are working at close to full capacity during this brief period. At some point after that first minute or so, the load balancer will typically drop one of the servers, re-discover it, add it back into the island, but that server will thereafter be unloaded - the performance meter registers 0% load. Note that this is despite the fact that the other server - still alive, is working flat out at 100%. Has anyone experienced this before ? I can't find anything in the archive. I'm especially perplexed since a lot of other people claim to have load-balancing and clustering working fine - after working from the same 'howto' as me. I would really appreciate help on this. Merry Christmas everyone. Tony. --- Dr Tony J Brooks Apama (UK) Ltd 17 Millers Yard Cambridge, UK Mobile : 07748 767 110 eMail : [EMAIL PROTECTED]