[okfn-help] further to load on us0

[Martin Keegan]

On Thu, 8 Oct 2009, John Bywater wrote:

> > Debian seems to want to prevent you from installing both prefork and
> > worker mode Apache simultaneously for separate instances, and I don't know
> > what of our software other than Django has assumptions about Apache's
> > process model.
> 
> Could always build a prefork Apache for KForge to use. Reverse proxying 
> KForge works so long as the reverse proxy also has the required modules 
> enabled. Building Apache for KForge isn't trivial because there are 
> quite a few separate downloads for the various modules. I've got some 
> code which automates (and documents) the sequence, if you're interested.

Ok, well, it's good to know we have that to fall back on, though obviously
even with the automation it's going to have a higher maintenance cost than
apt-get ...

> > Anyway, us0 is looking like less of a crimescene now. We can have the
> > "what should kforge look like" discussion later.
> 
> Good to hear. What do you think KForge should look like?

Where's the current roadmap?

I don't know what I think KForge should look like. I don't like the look
of the repeated boilerplate Apache config stanzas, and am wondering how we
expect it to be deployed (e.g., by an organisation which hosts KForge
instances on behalf of others?)

Anyway, that's a broader issue. What I want to know right now is how hard
it is to memoise the result of password lookups into some time-limited
cache, so we can reduce the number of calls out to postgres. 

Mk