Brian Moon wrote:
Just throwing out ideas. Thoughts?
Well, we have moved to MySQL Cluster for the bulk of the data that was
giving us this problem. We had major issues when anything on the front
page was not there. As for the cluster data, its just there or its not.
We don't use the pulled cache method for that anymore.
Mostly we use memcached now for proxy caching content and for caching
web services data from 3rd parties.
Thanks for all the responses ;) I updated the FAQ with more details
based on this. If it's incorrect please go edit.
I'll stop flooding the FAQ in another hour or two. I hope it's a good
enough start to motivate other folks into undoing my damage. It's time
I spent more time writing code elsewhere :)
I'm sure the OSCON BoF might produce plenty more interesting details for
the wiki (which unfortunately I missed this year). Someone should
check/post!
-Dormando
