Hello, I'm new to this list so let me introduce myself: My name is Stacy
Curl, I've been doing work for the past 3 months on JMX failover.
I've developed some software which monitors specific kinds of MBeans and
takes corrective action on the MBeans when the MBeans 'fail'. I've got
several layers of failover, the first 'corrective action' I take is to
simply invoke a restart method on the failed MBean, next if that fails to
work I reregister the offending MBean and start it up, next I kill the Agent
(my Agents are RMI activatable objects), next I reboot the machine.
None of this code is EJB specific but I do think that it can be a good
starting point to add clustering to JBoss.
My company looked into using EJB some time ago as a way to get 5 9's,
scalability, etc. but we unfortunately had to conclude that it was suitable,
mainly because we absolutely needed to create threads, due to comms stuff
(I've bludgeoned every BEA consultant I've met with this one). Leaving EJB
behind us meant that we needed to develop our own simple failover stuff,
hence my work. Why am I wittering on about this ? Well it seems that
clustering, high-availability, etc. don't just affect the 'narrow' problems
that EJB can provide solutions for, so perhaps with work (a lot of work)
JBoss could provide clustering outside of EJB too.
I'm just about to post my code up here. So you can have a look at it, rip it
apart, whatever.
Cheers,
Stacy.
