ugh, this does not work across a cluster. Only works on the member you are currently associated with. You have to fool around and ensure you hit each instance in the cluster with this.
DK On 6/9/05, Figy, Kam <[EMAIL PROTECTED]> wrote: > > <cfif StructKeyExists( URL, "reset" )> > <cfset StructClear( APPLICATION ) /> > </cfif> > > No downed cluster, app scope reset, and really handy. > > /kam > > -----Original Message----- > From: Douglas Knudsen [mailto:[EMAIL PROTECTED] > Sent: Thursday, June 09, 2005 1:56 PM > To: CF-Talk > Subject: Re: CFC's and limiting DB queries > > How do you get zero down time doing this? We have sticky sessions on, so > if > we take a member of the cluster down, we loose like 50% of my users > forcing > them to relogin to the other member, in this case we have two memebers > in > the cluster. > > DK > > On 6/9/05, Barney Boisvert <[EMAIL PROTECTED]> wrote: > > > > Application-scoped instances are a wonderful thing. And having a > > clustered environment makes them easier to deal with, not harder, > > because you can take a server out of the cluster, let it rebuild it's > > in-memory instances without load, and then reinject it. Repeat for > > the rest of the cluster, and you've got a zero-downtime upgrade. If > > you've got a large cluster, best to take half the servers down, > > upgrade them, then switch halves. As long as you don't need more than > > half your servers to deal with the load (which you shouldn't, because > > you should do upgrade under non-peak load, and most load cycles are > > more than 50% deltas), you can do it all in two steps. > > > > Particularly with larger applications, instantiating a large number of > > CFC instances on every request (rather than leaving them in the > > application scope across requests) can be a real bottleneck. Hundreds > > or thousands of milliseconds per request, if the instances need to do > > complex initialization. Almost always better to cache the instances. > > Just watch out for potential race conditions as multiple requests can > > be accessing the shared instances at any given time. > > > > cheers, > > barneyb > > > > On 6/9/05, Douglas Knudsen <[EMAIL PROTECTED]> wrote: > > > I'd suggest keeping gateways with objects or purposes, having a > > site-wide > > > one will certainly tend to grow to immense porpotions. This from a > bad > > > experience following me around where I work. Also, IMHO, stay away > from > > > application scoped cfc instances unless you really really need them. > > > They > > > are a PITA when you need to update the code, most notably when you > are > > in a > > > clustered environment. > > > > > > DK > > > > -- > > Barney Boisvert > > [EMAIL PROTECTED] > > 360.319.6145 > > http://www.barneyb.com/ > > > > Got Gmail? I have 50 invites. > > > > > > > > ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~| Logware (www.logware.us): a new and convenient web-based time tracking application. Start tracking and documenting hours spent on a project or with a client with Logware today. Try it for free with a 15 day trial account. http://www.houseoffusion.com/banners/view.cfm?bannerid=67 Message: http://www.houseoffusion.com/lists.cfm/link=i:4:209164 Archives: http://www.houseoffusion.com/cf_lists/threads.cfm/4 Subscription: http://www.houseoffusion.com/lists.cfm/link=s:4 Unsubscribe: http://www.houseoffusion.com/cf_lists/unsubscribe.cfm?user=89.70.4 Donations & Support: http://www.houseoffusion.com/tiny.cfm/54
