Sugg: Improve for mock (caching)

[Jeremy Katz]

On Thu, 2005-11-03 at 21:45 +0100, Enrico Scholz wrote:
> The speedup gained by caching is not (primarily) related to the network
> connection but to the rpm installation and disk-IO. On the fedora.us
> server we used caching also and the initial buildroot was created in
> <1min (afair).

I personally think that time is better spent investigating where this
time is spent because it will be far more generally useful than just
doing some caching for builds.  It would help users doing installs,
users installing updates and developers using mock

One thing that probably makes sense would be to set the rpm macros we
use at install-time in anaconda to change db4's configuration to be a
little bit less worried about ensuring everything is synced to disk.

> Local tests at that time showed, that the initial buildroot can be
> created in 20 seconds when a tmpfs will be used instead of a real disk
> with ext3.

I wish we had the RAM to do all builds on tmpfs or ramfs.  Alas, not
really practical at this point

Jeremy