[mythtv-users] Re: Why only numeric IP addresses work?

Maarten mythtv at ultratux.org
Wed Oct 13 21:45:15 UTC 2004

On Wednesday 13 October 2004 22:27, Scott Mazur wrote:
> On Wed, 13 Oct 2004 11:57:50 -0700, Jarod Wilson wrote
> > On Wednesday 13 October 2004 08:54, Grant Edwards wrote:
> > > On Wed, Oct 13, 2004 at 11:02:28AM -0400, Matt Mossholder wrote:

> > > I suppose.  To me something that never changes is static,
> > > regardless of where it's stored.
> >
> > I'm still not a fan of doing things that way. What happens if the
> > DHCP server croaks? All of a sudden, none of your servers have any
> > IP address (once their DHCP lease expires)...
> What happens if your myth server croaks?  All of the sudden you miss
> recording 9 episodes of I love Lucy and none of your TVs work.  So what?

True, but still, one could wonder whether it is more work to add many 
MAC-based 'static' entries in the DHCP servers config versus just configuring 
the few boxes that need non-changing addresses just with a static IP.
In any case, if we assume that the scope of this is broader than just mythtv 
backends, but all servers in general, I for my part want my servers to remain 
up and reachable no matter how much DHCP servers get killed on the LAN.
Cause most of my servers tend to have a 'slightly higher' uptime than some 
random W2k box or DSL modem that's assuming the role of DHCP server.

As a matter of fact, the few times I did configure a server through a dhcp 
entry I often gave it a ridiculously high leasetime, like 600 days or so, 
just to make sure it was not dependent on some botched DHCP server. ;-)

> Whether DHCP is a good or bad idea for anything is really wasted debate.
> Every network admin has a reason for doing things they way they do, and
> every network is different than any other.

Very true indeed.  Some people may even have their dhcp server running on I/F 
'lo' and hand out addresses there. But then again those people 
might want to rethink that procedure...  ;-) 



More information about the mythtv-users mailing list