I like keeping it to stuff that makes sense. I don't put in private network addresses unless I actually use them, which would just be the 192.168.x.x addresses provided by my DSL router, behind which I hide most of my systems. But for the present thread, I'm talking about the routable IP number for this host: treat.kosmanor.com is seen as 64.166.164.49.
I've got HOSTNAME=treat in /etc/conf.d/hostname, and the FQDN in /etc/postfix/main.cf. Meanwhile, I've stumbled on the Linux man page for uname(2). It explains some of this as lack of coverage from the standards. Use "man 2 uname" and look in the NOTES section. It seems this is an area that has not been well treated, and there's divergence among systems as well as among applications, let alone between them. Ugh. On Dec 2, 2007 7:12 AM, Etaoin Shrdlu <[EMAIL PROTECTED]> wrote: > On Sunday 2 December 2007, Mick wrote: > > > > Try adding the following line to /etc/hosts: > > > > > > a.b.c.d hostname.your.domain hostname > > > > of course, replacing a.b.c.d with your correct ip address. > > > > > > I don't know whether this is related to your problem, but it usually > > > solves the domainname: (none) problem. > > > > Is this meant to be the LAN private address, or the Internet address > > of the host? > > It depends. If you just need to solve the agetty banner problem which > prints "welcome to machinename.(none)", you can just use 127.0.0.1 or > eth0's address, and this always has solved all the problems for me until > now. Otherwise, you have to know what address the application uses to > identify the box. However, if in doubt, nothing stops you from adding > several lines that differ only in the ip address: > > 127.0.0.1 hostname.your.domain hostname > 10.0.0.1 hostname.your.domain hostname > 100.100.100.100 hostname.your.domain hostname > > etc. > > NOTE: I don't know whether this is the correct way to do things. Many > times it works, but other, less clumsy, ways probably exist (I simply > haven't had the need to search them until now). > -- > [EMAIL PROTECTED] mailing list > > -- Kevin O'Gorman, PhD
