[distcc] My big plans for distcc

[Daniel Kegel]

Victor Norman wrote:
> o Goals for the system:
>   o to pick best servers available, whether 1 engineer is compiling or 20 are.
 >   o to degrade gracefully under heavy loads.
 >   o fairness when heavily loaded.
 >   o to work with compiles dispatched from multiple
 > machines simultaneously (i.e., multiple "clients").

How 'bout as a first cut, just picking a random server?
That should meet the above requirements a little better than the current scheme,
especially when the server farm is uniform, and it's cheap.
I've implemented it; the patch is at

http://kegel.com/crosstool/crosstool-0.28-rc34/patches/distcc-2.16/randomhost.patch

>   o to work with our heterogeneous network of servers
> -- some very fast, some
>     multiprocessor, some very slow, some conditionally
> available.

Ixnay on the slow servers.  Best way to run a distcc cluster in my experience
is to never use slow machines as servers.

As for hetrogenous servers, I've run clusters that contain linux, MacOSX, and Cygwin
boxes.   I did patch distccd a bit, but only because I wanted the
compilers for all three build systems to be visible from all systems,
so I needed distccd to ignore the first couple components of the compiler path.
This is up and running, and is part of the devel version of http://kegel.com/crosstool

>   o an "availability server daemon" will run on a
> single machine and will
>     write to the hosts file, adding machines that
> become available, and removing machines that are not
> available anymore.  And it will write the
> load-averages  into the new hosts config file.  It
> gets this informationi by periodically receiving
> messages from each machine that is configured to
> participate in the compilation system. 
>   o All machines in the compilation farm will run a
> small client daemon to
>     communicate with this server daemon, so that the
> machine's load-average 
>     and availability status is updated periodically. 
> A machine's
>     availability may be determined by its
> load-average, or when its
>     screensaver is running or not, etc.
>   o we have a prototype for this already -- it is
> basic socket programming,
>     mostly. 

IMHO that's overkill.  However, go ahead and post your code, it might be interesting.
I keep trying to figure out a way to do something like that with it
being useful.  Maybe you succeeded.

- Dan