On Thu, Dec 9, 2010 at 11:45 AM, Russell Keith-Magee
<[email protected]> wrote:
>
> During the original multi-db design process, we considered allowing
> for the definition of dependencies, but abandoned the idea because of
> the complexity required to get automated synchronization correct in
> every case. For example, a master/slave setup only requires syncdb to
> be run on the master -- how do you declare this in DATABASES? What
> about multi-master configurations? Sharded configurations?
>

Hi Russ

I just wanted to give you a customer viewpoint of how we deal with this.

I currently run a production SSO solution for a global enterprise that
is built around django, and use multi master replication to have
replicas in different data centres around the world, with slave db
servers in each location for reads, all built around django's multi db
support.

When we have schema changes, we have a complicated sequence of moving
requests to a single data centre, halting DB replication from the
inactive data centre. We then update the schema on the inactive
master, re-activate replication, and move requests to the previously
inactive DC. We then run syncdb against one of the masters, in order
to add non schema changes (eg, adding a new ContentType).

This is a complex procedure, and we don't expect django to do this
automagically. We're quite happy with the tools django gives us, like
SQL generation for our migrations, etc.

Cheers

Tom

-- 
You received this message because you are subscribed to the Google Groups 
"Django developers" group.
To post to this group, send email to [email protected].
To unsubscribe from this group, send email to 
[email protected].
For more options, visit this group at 
http://groups.google.com/group/django-developers?hl=en.

Reply via email to