The Future of Performance: Scaling

Bradley Baetz bbaetz at acm.org
Wed Dec 31 04:16:20 UTC 2008


On Fri, Dec 19, 2008 at 9:46 PM, Max Kanat-Alexander
<mkanat at bugzilla.org> wrote:
>        So, I was reading this:
>
>        http://www.codinghorror.com/blog/archives/001198.html
>
>        And it was a good reminder of something that I've known
> for a long time, but sometimes had forgotten--the future of performance
> is, by and large, *scaling*, not optimizing. That is, tweaking our code
> is not the important thing--what's important is that people should be
> able to throw *more hardware* at Bugzilla to make it faster, if they
> need to.

In general, we scale read-only pretty well. The only issue I'm aware
of that prevents multiple web front ends is data/params and
localconfig (there is a bug on that IIRC). The dependency graphs
probably also require a shared datadir - not sure if theres a bug for
that one.

The shadowdb stuff pushes db reads off, but Bugzilla doesn't have a
clear goal as to how out of date data can be - if we relaxed the rules
a bit then some more queries could possibly move to the replication
db. However, I don't think that thats much of an issue given the
typical read:write ratio.

Bradley



More information about the developers mailing list