Shiraz wrote:

> Multipool HA & LB Algorithm are used to manage multiple DB servers.
> >
> >In General the maximum number of connection that an weblogic instance
> >will
> >be less then execute thread count.(Executethread count -3).(Application
> >should not create it's own thread).

> Then how do you achieve some scheduled asynchronus processing for e.g a batch
> job that runs after every hr or some frequency.
> You wouldnt want to tie your code to weblogic API to get a thread from weblogic
> maintanted threadpool. I guess Timer addition can be used for this but if i am
> correct its in the J2EE 1.4 spec.

Lots of stuff may be in the jdbc spec that would hurt overall performance of
a multiuser system like WLS. You should use the WLS API for asynchronous
services. Otherwise you risk an unbounded parasitic drain on the JVM resources
that WLS needs for serving all users. If you are confident in totally in control
about what the WLS JVM is being asked to do, and you find it better to roll you
own threads, so be it, but if the WLS is for a broader use than for the purpose of
your spawned threads, then you may be inadvertantly robbing Peter to pay Paul.


> -shiraz