PGday.eu
Sign at entrance of PGday.
I’ve been in Prato for the last two days (well, actually, only two nights) for the first official european PGday! PGday is a gathering of PostgreSQL geeks from europe and beyond. For those that might be wondering what PostgreSQL is, it is the leading open-source database software. Whereas most people might know mysql from their website hosters, PostgreSQL is used by anyone with serious database needs (say, databases with sizes in gigabytes to terrabytes) and increasingly is the choice for enterprises using oracle or informix before.
So i’m once more visitting the country of icecream, pasta, using the car horn instead of the blinkers and a requirement to put your ID on file before being allowed to go on the internet. The venue is Monash university, in a nice old building not far from the local castle. I haven’t seen that much from the city though, as the hotel is only 5 minutes walking distance. On my first night, i had the company of two mosquito’s in my room. I didn’t get much sleep until after i terminated those.
PGday is like a warm bath, the PostgreSQL community is one of those communities that just feels good. I see a lot of familiar faces, people i already met at fosdem 2008. But a lot of new faces too. The european PGday is a combination with the Italian PGday. In fact, last year it was called the Italian PGday. But it had such attraction to non-Italians, plus the fact that in february the European user group was formed, that this year it is truly european. The intention is to make it a yearly event.
We are now on the beginning of day 2, with a lecture by David Fetter about one of the many great new features: recursive queries. Yesterday, there were already some lectures on more new features: windowing functions among them. They’re great for time-series analysis, among other uses. Later today there will be a talk about in-place upgrades, which I hope promises to enable upgrades between major version numbers (eg. 8.2 to 8.3). Upgrading to a new major version is now more often than not delayed (or even skipped) because the down-time that comes with a dump-restore cycle on big databases is not acceptable in mission-critical databases.
Next to those DBA-targetted talks, there is a track for developers too, that starts with the next talk (the one i’m waiting for now in one of the nice bars in the building). What I personally would have like to have seen is sort of an introduction to the PostgreSQL code base. I haven’t yet gotten my hands dirty on any serious PostgreSQL development, apart from some small tours in the codebase to check why a certain feature did this or that.
Back of program leaflet.
One thing that became clear from the keynote speech is that the development model that has been in use for the 10 years that PostgreSQL is in existence so far is not suited for the continued growth of the project. As more and more developers and users are getting involved with the development process, a new model is needed. Changes are being made currently to accommodate the growth of the project. And this is perhaps a way for me to start a career as a PostgreSQL developer too: there is a need for patch-reviewers. Initially, one would review the patch on the basis of some basic rules (i’d guess mostly syntax and style). A great way to get to know the various parts of the codebase from up close.
I’ve been a PostgreSQL user and DBA for a while now, and recently I decided to make it one of 2 specialisations of my company Sonologic. As you can see on the leaflet on the right, Sonologic is one of the sponsors of this event. It is a small chance to do something back for all those people that created and still maintain PostgreSQL. Hopefully, in the future I will be able to do even more!
At least the meeting has instilled some fresh enthousiasm in getting something like a DBPUG off the ground: a Dutch-Belgian PostgreSQL Usergroup. At the very least, i’ll be manning a PostgreSQL booth at T-Dose together with one or two other Dutch PostgreSQL enthousiasts. I must nog forget to ask whether I can take some leftover shirts and other swag to sell. But now, the talk about custom indexing with GiST indices is starting so time to log-off!
