[dev] Multiple SQL Server Support

[Michael Cochrane]

After a few seconds thought, the obvious place to implelement this seemed to be 
a PEAR::DB class that deals with two servers, either directly or indirectly via 
two (or multiple) PEAR::DB ojbects. 

The second way allows for different DB servers (eg MySql, SQL server). And 
could also implement options like load balancing (choose a server at random) or 
fall over support (work through a list until a working server if found)

Implmenting it as a PEAD::DB class means that no changes are needed to the 
Horde framework nor any other applications that make use of the PEAR::DB 
classes to connect to databases.

Just my thoughs.... 

- Mike :-)


Quoting James <james@james-web.net>:

> Idea: 
> Redundant SQL Server Support.  If the 'primary' SQL server can't be
> contacted, Horde and whatever configured applications switches over to a
> 'secondary' SQL server.  This would probably require a redo of a lot of
> ..conf files, and some work inside of the core horde classes.
> 
> But I think it's a large benefit as it provides failover survivability
> in case one SQL server is hit by a meteor or something.
> 
> I'm currently utilizing this in non-Horde related in-house application,
> and was wonder your thoughts on its feasability inside of Horde.  Would
> this be a easy to moderate project, or a major to impossible project?
> 
> - James, fiendishly planning
> 
> 
> -- 
> Horde developers mailing list
> Frequently Asked Questions: http://horde.org/faq/
> To unsubscribe, mail: dev-unsubscribe@lists.horde.org




-------------------------------------------------------------------------
This mail sent from Mike's CVS HEAD install of IMP: http://horde.org/imp/