[distcc] distcc at large sites. Patch to get hosts file from /etc/distcc/$dnsdomain/hosts

[Daniel Kegel]

Martin Pool wrote:
> It seems like this approximates the other proposed behaviour of having
> distcc.$SITE.company.com have A records for all the servers, and then
> doing locking etc by ip address.  It has the advantage of requiring
> less cooperation from the DNS administrators, but the disadvantage of
> being a bit more of a special purpose hack.

I like to think of it as "appropriate technology".  It's
darn hard to get the attention of the IT staff at
most large companies, so minimizing their involvement
is a key to success in many environments.

Also, the idea you mention has a problem when the set of A
records is larger than 512 bytes, doesn't it?  (Or is RFC2671
universally deployed already?)  That would make it hard
to use for large clusters.

I've polished and debugged the patch; the new version is at the same URL,
   http://kegel.com/distcc-domain.patch
It no longer has any appreciable overhead beyond an extra stat
when the feature is not in use, so it's safe to always enable.
I haven't deployed it yet, but I'll try to do that this week.

Incidentally, my init.d script that starts up distccd will read
the list of allowed networks from /etc/distcc/$domain/hosts.allow,
so the idea of a $domain subdir for /etc/distcc
seems to be more than just a special purpose hack.
I suspect it will be generally useful at large multisite companies.
- Dan