Chris Betts wrote:
> Hi Folks,
>
> it looks like I've made a blunder; I had thought partitions were
> 'pluggable' back end data stores.
No you are right!
> However it seems that in fact they're just named areas of the default
> database implementation.
Well when you have an entry store it needs to be attached to the
namespace at some point right? Hence ContextPartition.
> This is a bit of a pain, as what I want to do is hive off part of
> the namespace and handle it myself using my own code. I *could* just
> write a jndi provider and use the SEDA front end, but then I'd have to
> do the namespace partitioning myself, and I'd prefer the Root Nexus to
> do that.
Yeah totally the root nexus will do that for you. Don't mess with
writing a new JNDI provider and using SEDA in general. FYI MINA is the
new network protocol API we're using. It's much better.
> So what's the best way forward? Should I use the Interceptor framework?
See if you can implement a ContextPartition.
> ... I'm still getting my head around all this :-)
Np :)
Alex
|