This creates an interesting predicament. One of my expectations is that if I give a machine a name and it gets an address by DHCP. I should be able to use that machine’s given name to find it by DNS. With regards to leases, I have been spoiled by firewalls where I can just click on a dynamic lease and convert it to a static lease.
I should probably explain why am trying to use UCS. My end goal is to replace NIS. In many ways I love NIS because one of the simplest, time efficient authentication systems out there. It sucks because security is nonexistent. Still though like the post-it note on the keyboard, it makes logging into a system very end user and admin friendly.
I got to the DNS question because I’m trying to understand what I need to do to have a single set of login credentials so people a using SSH, VPNs, NFS can work across multiple servers. Truth be told, I’m very very close to using SSH to push around fragments of a various files like a password file because that’s almost as simple as NIS and it meets my functional needs. I also know it’s not a wise approach but it’s driven by frustration that authentication is moved to something that is secure with a low usability factor.
UCS looks really good in a large scale. It probably makes things much easier to manage than other systems out there (<cough>FreeIPA</cough>). I’m just not confident I can figure out enough about the system to make it work for me.
Maybe it’s time to hire you guys to set it up and it as an unmanageable (by us) black box.
