Paul J Stevens wrote:
> 
> Very interested! Though this belongs on the dbmail-dev list really. One
> of my plans post-3.0 is support for user-based sharding. Not the same as
> what you propose, but there may be enough overlap that we need to make
> sure we don't make life hard on each other :-)
> 

My bad - I couldn't decide where to post this and opted for here for user
feedback.

That said, I'm not sure if you noticed but I committed my changes to a
branch on GitHub here:
https://github.com/chrisboulton/dbmail/tree/masterslave

In terms of what's left to do:

* Testing. I'm not running DBMail in production yet anywhere, so all I can
do is test locally, using the app as best as I can. I don't know all of the
code paths (having only looked at DBMail for the first time yesterday). What
I need to make sure of is that there's absolutely no writes going to the
slaves (which should be read-only), and that there aren't consistency
issues.

* Look at the CURRENT_TIMESTAMP references. I'm pretty sure these are safe
with MySQL and statement based replication, but not sure about Postgres,
Oracle, etc.

I by no means consider myself a C programmer, so I'm also looking for
feedback on my implementation.


Paul J Stevens wrote:
> 
> Being able to specify different storage backends for blobs would be
> awesome! One gotcha you need to be aware of: IMAP-SEARCH. But that could
> be solved by adding support for SOLR/Lucene.
> 

Glad to see we're on exactly the same page (I was thinking Elastic Search
which is built on Lucene). Either way, pluggable modules here would be
great.
-- 
View this message in context: 
http://old.nabble.com/Interest-in-read-write-database-pools--tp31468970p31474527.html
Sent from the dbmail users mailing list archive at Nabble.com.

_______________________________________________
DBmail mailing list
[email protected]
http://mailman.fastxs.nl/cgi-bin/mailman/listinfo/dbmail

Reply via email to