Brian’s Postgres Post is becoming an Internet Landmark June 2, 2007 | 06:31 pm

I just bumped into Brian’s post on Postgres’s performance while researching running JIRA on Postgres.

Checking out the web server statistics, it looks like that page had about 1000 hits on it, and only the RSS feed and robots.txt exceed it.

Congratulations, Brian!


  • Brian

    Actually, the argument that I was prematurely optimizing now has evidence behind it. We have tables running into the tens of millions of rows, with no segmentation, and little optimization, that are acceptably fast (i.e. we didn’t notice how big they’d gotten).

    I really need to revist that topic…

  • Robert

    The article also implies that you need to do a lot of smart work to get Postgres to be fast, which probably isn’t fair.

  • Brian

    You’re right. Although I will comment that it does require some smart work to get it working. Reading the chapter on configuration is basically a requirement, as is setting up some sort of regularly scheduled maintainance (man cron).

    On the other hand, this is another pattern I’ve seen repeated over and over again- the requirement that you put in a little brain work early to save a lot of brain work later. Given all the wailing and nashing of teeth I’ve seen on the internet over MySql syncrhonization issues and transactional issues, I seem to have avoided a lot of work downstream in return for a little work upstream. A similiar pattern shows up with Haskell- you pay an upfront brain work cost learning monadic style and functional programming, but it pays dividends in a big way with parallel programming.

  • Pingback: Enfranchised Mind » Sweet, Sweet Link Lovin’ from Atlassian