23550 Posts in 19336 Topics by 2890 members
|Go to End||Next >|
31 July 2009 at 4:48pm
Heh, yeah I went hunting for it and realised that it was handled in Core.php
Why not bundle it into ManifestBuilder as a static method?
2 August 2009 at 9:25am
Have you guys thought about building out a memcache-d requirement and using memcache to store that huge manifest of file names/positions and info?
2 August 2009 at 6:27pm
> Have you guys thought about building out a memcache-d requirement and using memcache to store that huge manifest of file names/positions and info?
Don't you think that using memcache is a bit exaggeration for a CMS website?
5 September 2009 at 5:16pm
Sure...exaggeration if we're talking about <1MM pv/month. I'm talking about high-volume, load-balanced use.
6 September 2009 at 9:26pm
Most CMS websites are not "<1MM pv/month" ones. That's why making memcached obligatory is an exaggeration. Making it optional is something else...
7 September 2009 at 5:12pm
Maybe required was the wrong choice, but if it were an insurmountable performance issue it's not far fetched IMO.
I think that any implementation worth its salt that uses any 3rd party "requirement" (in this case memcache-d) would be crap unless it had a fall back in case of unavailabilty. Using memcache-d, if you wrap a DB class with the memcache correctly, you code it to default to hitting the DB directly if memcache is not available.
I'm not sure what you're getting at...was it my semantics? Have I been sufficiently "called out" with "quotations"?
I was trying to explain that, if memcache were available and you used it on the sections that were causing bottlenecks (provided it's not constantly changing data ala your user profile) it makes a geometric difference in performance.
FWIW, I've hacked in memcache support for select DB calls (no pun) and have witnessed a 3.5x ability to serve non-static data with a 30 second memcache expiration (or on change expiration rules)...
19 October 2009 at 5:20am
Ok, than you didn't want memcached as a requirement but as an option. I agree of course ;)
19 October 2009 at 7:45pm
>FWIW, I've hacked in memcache support for select DB calls (no pun) and have
>witnessed a 3.5x ability to serve non-static data with a 30 second memcache
>expiration (or on change expiration rules)...
I'm keen on learning what hack you did and how you went about it. ;-)
Having some issues with a pretty large site a.t.m..... Using APC to store pre-rendered HTML , not the nicest of options... ;-(
|Go to Top||Next >|