The whole discussion of Reactor and memory leaks started nearly 6
months ago and it's cropped up several times with people saying "I
have reactor in one version of the site and not in the other and the
one with Reactor blows up every {6 to 48} hours." I'm starting to
think this is a function of usage and/or developer "error" (error in
the sense that the dev is doing something that's not quite kosher and
different than the plenty of people who DON'T have this issue). I'm
not coming down on anyone, but I think we might want to switch this
conversation from "Reactor is causing memory leaks" to "what's
different between those that have the issue and those that don't have
the issue?"
Even Doug was having issues with ReactorBlog and a memory leak that
would down his instance every n-hours... so Dave, how are you using
Reactor and it's generated objects? Are you using Records as "beans"
and passing them around? Are you using the TOs instead?
Do you keep ReactorFactory in the application scope?
I think we should be focusing on finding the differences between one
implementation that bombs and one that doesn't...
Anyone?
J
------------------------------------------------
Jared C. Rypka-Hauer
Continuum Media Group LLC
http://www.web-relevant.com
Member, Team Macromedia - ColdFusion
"That which does not kill me makes me stranger." - Yonah Schmeidler
On Dec 12, 2006, at 12:20 PM, Dave Shuck wrote:
For what it is worth, we exclusively use Mach-II/Reactor on
InstantSpot. Our memory has stayed on a completely flat line since
we launched it in early September, with only slight increases
during heavy traffic times as we expected, but is subsequently
released as it should. We have only restarted the server once and
it was due to an unrelated ColdFusion issue regarding mail queues.
~Dave
On 12/12/06, Messer, Jon <[EMAIL PROTECTED]> wrote:
I have this exact problem although less frequently. I am using
ModelGlue
Unity with reactor and running on CF with the internal JRun. The
process
memory maxes out occasionally and the server stops responding. I
haven't
tracked it down yet, and I can't replicate it under controlled
circumstances.
I installed the 10 day demo of Fusion Reactor to see if that would
show
anything. When it happened it was a single request that just chews up
the memory (but not the CPU's), if I kill that thread the server
recovers. If I then rerun the request that caused the problem it
works.
So it's very intermittent. I wasn't sure if it was my code, reactor or
ModelGlue.
I haven't solved it yet, though sorry.
-- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --
Reactor for ColdFusion Mailing List
[email protected]
Archives at: http://www.mail-archive.com/reactor%40doughughes.net/
-- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --
