JRun has a cluster manager service. When a server is offline or stops responding to requests, it's dropped from the cluster. If you attempt to access something in the session scope and the server is unable to access the session, it fails over to the other cluster member... but only if you have fail over turned on (which you should). For the most part, nothing much short of a crash, stop, or restart will remove an instance from the cluster as far as I know. This is one of the reasons you balance the instances. When it is restarted or started for the first time, the cluster manager adds it to the cluster for you.
~Simon Simon Horwith CIO, AboutWeb - http://www.aboutweb.com Editor-in-Chief, ColdFusion Developers Journal Member of Team Macromedia Macromedia Certified Master Instructor Blog - http://www.horwith.com Brandon Harper wrote: >I've been setting-up/testing a simple JRun cluster with instances of >ColdFusion MX 7.0 and one thing which does not seem to be documented >well about JRun is what events signal adding/dropping instances from a >cluster, what errors it can or can't recover from by switching to >another instance in the cluster, etc. If someone could pass on a link >about this that I haven't been able to find, that would be great. >Otherwise I have a laundry list of questions below... > >Here is the setup I'm using (all are Windows 2003 servers): > >1 Web server connected to 1 JRun Cluster > >JRun Cluster contains 2 servers > >Each JRun Server contains 2 instances of a given application, so there >are basically 4 instances in the cluster total. > >No load balancing between the web server and the JRun servers or any >of that sort of stuff yet, it's just a basic cluster I'm setting-up >for a product beta. > > >Here are my specific questions (that is, if there isn't anywhere I can >read about this stuff)... > >1. Other than stopping an instance, what events/exceptions would >trigger a JRun application instance to be dropped from the cluster? > >2. What events does using the JRun drivers pointing to a cluster (as >in not using a load balancer) help catch gracefully? So far it looks >like only when an instance has stopped. For instance I've got several >"cannot connect to JRun" errors when I've intentionally made a server >busy. I haven't tried shutting down a server completely then trying >it out yet-- does it catch that? > >3. How does an instance which has been restarted or otherwise dropped >from the cluster make it's way back into the cluster? I've tried a >lot of scenarios with this in particular, and can't figure it out. It >just seems to eventually reappear, but it takes a very long time for >that to happen. > >Mostly I'm just trying to get a good feel for how well or not this >scenario works so that I can troubleshoot potential problems with it >in the future. I'd greatly appreciate any advice or any useful links >you could send my way. I've read tons and tons of stuff about >clustering in JRun, but nothing specifically about how JRun clustering >behaves/reacts in different scenarios. > >Thanks, > >- Brandon > > > > ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~| Logware (www.logware.us): a new and convenient web-based time tracking application. Start tracking and documenting hours spent on a project or with a client with Logware today. Try it for free with a 15 day trial account. http://www.houseoffusion.com/banners/view.cfm?bannerid=67 Message: http://www.houseoffusion.com/lists.cfm/link=i:4:202454 Archives: http://www.houseoffusion.com/cf_lists/threads.cfm/4 Subscription: http://www.houseoffusion.com/lists.cfm/link=s:4 Unsubscribe: http://www.houseoffusion.com/cf_lists/unsubscribe.cfm?user=11502.10531.4 Donations & Support: http://www.houseoffusion.com/tiny.cfm/54
