James, JB> On 10/3/07, Andrey Khavryuchenko <[EMAIL PROTECTED]> wrote: >> The only thing django app should know is that there are two different >> *kinds* of servers: read-only and whole access. It doesn't matter how many >> servers you have - you may well have pool of masters and pool of slaves, it >> doesn't matter.
JB> For *your* case. In my case I need no replication nor clustering: project is fairly small now. My blog post and question here where to that point just a mind excercise. JB> But remember that Django is expected to be generic and broadly usable, JB> which means that actually the decorators would have to support, say, JB> people who are sharding and want to have this view write to that JB> shard, and that view read from another, etc., etc., which means that JB> your decorator scheme basically comes down, in the end, to hard-coding JB> database or cluster or shard names alongside each view. Does django orm support sharding now? No, as far as I know. So let's solve problems as they come. JB> And again, I'd argue that this isn't something the application layer JB> needs to know or should have to know. >> And, completely unrelated to above, some companies made *business* decision >> to use MySQL Replication widely. So if django doesn't support it, no big >> projects using django. JB> You mean no projects using MySQL Replication, Did I? I think I've said exactly opposite. JB> which is very, very, VERY far from "no big projects" -- there are JB> plenty of big projects using Django, they're just using smarter DB JB> setups. Personally, I think MySQL's standard attitude of "well, just JB> completely rewrite your application when you need more than one JB> database" is the thing that's not suitable for big projects (where you JB> need to be thinking about, and planning for, configurable scaling JB> across multiple machines from the very start, not pausing and JB> rewriting when it's necessary); anybody who paid a consultant to get JB> that recommendation should ask for their money back. Perhaps. The company I've had in mind is not my company, instead it is a leading search engine player in other country. And I won't like to discuss *that* stuff further. I am interested in technical discussion, and not "you shouldn't use this (mysql replication), you should use that (mysql clustering)" - I know at least two people that are interested specifically in using mysql replication in django projects. -- Andrey V Khavryuchenko - http://a.khavr.com/ Django NewGate - http://www.kds.com.ua/djiggit/ Chytach - http://www.chytach.com/ --~--~---------~--~----~------------~-------~--~----~ You received this message because you are subscribed to the Google Groups "Django users" group. To post to this group, send email to django-users@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/django-users?hl=en -~----------~----~----~----~------~----~------~--~---
