Re: Persisting a one-off hostname change
On Tue, Feb 28, 2012 at 1:08 AM, Jason Heeris <jason.heeris@gmail.com> wrote:
> On 28 February 2012 13:40, Tom H <tomh0665@gmail.com> wrote:
>>
>> Everywhere that I've worked the hostnames have had something to
>> indicate its purpose and its location.
>
> I don't think this reasoning can be applied here though. There will be
> dozens of identical devices plugged into the network, and hundreds in
> total (but not all plugged in at the same time). They can't all have
> the same name, and the IP address is useless as an identifier since
> the physical location could change daily until it goes out the door.
I was just telling you what I'm used to. In my current job, we're
creeping up to 8,000 RHEL and Solaris servers. Random names would be a
real pain but we'd get used to it, I guess (!), if policy changed and
we implemented it. We have a db that we can query for every hardware
and software detail of every server, but we couldn't shout out across
to the next desk "they've opened yet another ticket for "<hostname>".
What I meant by physical location wasn't "this rack and this slot" but
in which data center the server's housed (or server room for a smaller
company).
>> The "mv /etc/rc.local.final /etc/rc.local" would be the last line in
>> the first-boot's "/etc/rc.local" so that your "/dev/random" hostname
>> stuff only runs once. I've just tried it and it was OK.
>
> Oh, right, that makes sense now.
My fault. I should started with "cp /etc/rc.local /etc/rc.local.final".
Reply to: