FYI: just wrote that little thingy during the sessions at ApacheCon -
hopefully this and switching to Derby will improve the stability of
the Modeler prefs database.
Andrus
Begin forwarded message:
From: Andrus Adamchik <[EMAIL PROTECTED]>
Date: June 29, 2006 7:01:47 PM GMT+01:00
To: <[email protected]>
Subject: Re: Derby as desktop database
Reply-To: <[email protected]>
This is the later - I don't want the users to start the server.
BTW, I just prototyped a simple distributed peer-to-peer DataSource
[1]. In my tests I can run about 7 parallel processes on my MacBook
laptop that access the same database non-stop (each doing a select,
then insert, then going to sleep for 2 seconds). When I started
more processes, timeouts and errors start to occur, but playing
with the DataSource parameters can increase the concurrency.
Not ideal but works. I may clean this up in the future, removing
Cayenne EventManager dependency, coding it straight to the
JavaGroups API.
Comments?
Andrus
[1] http://svn.apache.org/repos/asf/incubator/cayenne/sandbox/derby-
pool/
On Jun 29, 2006, at 6:25 PM, Mike Matrigali wrote:
Andrus Adamchik wrote:
Not a problem :-)
Yeah, I am looking for a solution that would work with a bunch
of independent applications.
Andrus
Just to understand is the reason network server solution is not ok
performance? Or is the problem that you don't want to have to
start a server up just in case one of the independent apps wants
it?
