In Seattle we hosted a developer meeting for Drizzle, Memcached, and Gearman.
The highlights for Drizzle?
There were a number of lively discussions. One included vendor support for the storage engine vendors/folks who are bundling the microkernel for their own releases (AKA how do we make this simple). A bit of time was spent looking at how to support different models of replication. There was a session led by Robert Hodges, of Continuent, about different models and what we should be doing to support these.
Clustering came up several times, and we went over the new sharding capabilities in libdrizzle, and talked about "Hydra" which is the idea for "redundant arrays of isolated databases".
The new Information Schema system by Padraig was brought up and we established the roadmap for removing the final bits of the old system (which will reduce the number of code paths).
A long discussion occurred around our need to verify more applications with Drizzle. A link can be found here: http://drizzle.org/wiki/Applications_to_be_verified. The Drizzle JDBC Driver is in use but it could use more testing (and more funding). Its ability to cross connect MySQL and Drizzle should provide it a larger audience (especially for anyone who is dealing with application licensing hell).
On Memcached we spent time going over the current state of drivers and what we need to be doing to get the server better tested. I went over some of the work I have been doing for embedded memcached. Matt Ingenthron brought up the concept of creating a matrix of known memcached server and listing out their compatibility with the protocol specification. Dustin Sallings talked about working on a driver for support of the Java caching JRE that other caching services use.
For Gearman there was talk of the current drivers, which are in good shape. We talked about the features on the roadmap and we spent time looking what we need to be doing to make sure that the project is well maintainable. Eric Day went through an example run of a job to a worker. We talked about the need for a better message/description of what Gearman is.
The next large get together for Drizzle will be in January. No venue has been chosen. We did decide to open the meeting up to more people. So like the Drizzle Developer day back in April we would expect to get around ~100 (which would be very dependent on location). Even though we had a dozen developers at the meeting, which included a large number of the core developers, both Sun and non-Sun, there were a lot of additional folks who commented that they had wished we had found a larger space so that more folks could be there. So sometime in the next couple of months we will look for a space on the West Coast.
There was a collective talk around Buildbot and Hudson. All projects share resources now, but we could be doing a better job of co-ordinating this. Mike Shadle brought up what we had, Dustin Sallings talked about his work with buildbot, and Lee Beiber went over our current testing plans (and what we are doing for expansion).
A number of the Drizzle developers will be at SQLCamp in Portlland, http://opensqlcamp.org/. We expect a good attendance and will be soliciting talks for what people would like to hear.