Hey all, I'm in the process of building in support for specifying a different database server to send all write operations to (traditional master/slave replicated environment) among other things at the moment.
I think this is particularly useful for DBmail as there's going to be components that are particularly write heavy (receiving mail, deleting mail, changing statuses etc) but there's a tonne more that's read heavy, and it'd be great to separate those off at the database level. Further down the track, it'd be nice to specify a pool of read servers (though that could just be loadbalanced at the moment with something like LVS) for querying. Is anyone interested in this? If so, I'll see what I can do to have it all finished off, and working with an experimental patch. I'll post it here for testing, and hopefully future inclusion. (The other thing I'll be working on shortly which some may be interested in, is the ability to build out modules for storing email parts. This would allow you to shift message content off to something like OpenStack Object Storage, or HDFS, which I think are tools better suited for mass storage than a database, and would mean the database scales a lot easier (and uses less disk space) along side a separated component for message storage) Regards, Chris Boulton Lead Engineer BigCommerce / Interspire Web: http://www.bigcommerce.com Web: http://www.interspire.com
_______________________________________________ DBmail mailing list [email protected] http://mailman.fastxs.nl/cgi-bin/mailman/listinfo/dbmail
