So, I just looked into how I handle this internally.
By default, the server uses the hostname contained in the HTTP request.
If this name disagrees with the hostname in the certificate, then the
certificate name is used.
This is the hostname passed to all XMLRPC calls. Most XMLRPC calls
don''t include facts, so I can''t use the facter-derived
hostname except
for in the processing of the ''getconfig'' method, which
requires a fact
list as an argument.
For that one method, if a client is provided (which would always be the
case when networked), then that name is used, meaning we''re usually
getting the certificate name. If no name is provided, then we use the
Facter name.
<...10 minutes later...>
Okay, the bit is flipped, and there''s a test in place to, um, make sure
I don''t do this again, I guess. Commit #1433.
I won''t be putting out a release at least until I fix the OS X
puppetmasterd bug, but I guess let me know if this is urgent. I''m
going
to be out of town again for a family reunion from Thursday to Sunday, so
if this is critical to anyone, let me know now.
Otherwise, I''ll spend most of the week working on some refactoring that
a friend is helping with:
http://reductivelabs.com/cgi-bin/puppet.cgi/wiki/ModelsAndImplementations
I think getting this to work will have a huge impact on people''s
ability
to contribute code, and it will also enable developers to declare
support for specific platforms, akin to how Facter can confine fact
resolutions to specific platforms.
--
Get forgiveness now -- tomorrow you may no longer feel guilty.
---------------------------------------------------------------------
Luke Kanies | http://reductivelabs.com | http://madstop.com
