Industrial PostgreSQL Thursday 14:40 New York
GitHub: moench-tegeder Company website: www.2ndquadrant.com
I’m Christoph, and I work as a consultant for 2ndQuadrant. I’m based in Karlsruhe, south-western Germany. I help our customers with all aspects of PostgreSQL - planning, configuration, performance and SQL tuning, application development. One of my primary interests is enabling organizations to use PostgreSQL efficiently: to get the most out of our favourite open-source database, you need much more than a few configuration settings.
I do a lot of work on-site with our customers, so I travel a lot all over Germany and Europe. As I work with rather diverse software (whatever the customer uses), I’ve sent in a few patches and bug reports for software in the larger PostgreSQL ecosystem.
I’ve been to most of the pgconf.eu conferences (I missed a few), and to all or almost all FOSDEM PGDays. I’ve also been a speaker on some of these events.
On the surface, this is a typical user story: we help a customer to introduce PostgreSQL into their service landscape. But this customer is not the typical customer we often hear about at the conferences: not only that they are a huge company, but they have been using IT for longer than PostgreSQL exists, and the company itself pre-dates IT by several decades. The size and history of this customer create some unique challenges - challenges not found in “fast moving” IT (or “mainly IT”) companies. I have learned a lot while working with this customer, and I want to share my experiences: with those who might find themselves in a similar situation, but also with yet-to-become-PostgreSQL users.
I will talk about processes, culture and communications, but also about tools and service architecture, so there’s content for many different people: DBAs and DevOps engineers, service managers and architects. I do not have much on PostgreSQL internals or deep hacking in this talk; the not-so-technical attendees should be very comfortable in my talk.
My presentation is rather light on the technical side - you don’t have to be a developer to follow the contents. Attendees should have some experience or at least an interest in managing a fleet of servers - not necessarily PostgreSQL servers, or huge data centers.
The Partitioning Improvements are great - especially the index improvements. Having PRIMARY KEYs and FOREIGN KEYs on partitioned tables makes them first class tables and allows to use partitioning in much more cases than before.
The WINDOW clause improvements (full SQL:2011 support) looks like a small deal, but it will really help on some statements - the workarounds we had to use before were quite ugly.
Gülçin’s “PostgreSQL is not your traditional SQL database” looks very interesting - PostgreSQL has so much more to offer than some commercial products, and users could have a much easier job if they knew about all the special features. And there’s Marco’s and Rainier’s “Partition a 10 TB table”: having so much data has it’s very own challenges; and partitioning that in a running application is far from easy. I’d like to see how they did that...