Linode resolvers vs own resolver
I've always followed the official Linode guide to use the two DNS resolvers listed in 'Remote Access' in /etc/resolv.conf
I built a new Linode yesterday, and the resolver was set to 192.168.1.1. I assume this is pointing at an internal daemon on the Linode.
Is there an advantage to either one? Perhaps the second method caches?
My guess is that it's a misconfiguration, either 1.) An internal setting that leaked out when Linode built the disk image, 2.) A DHCP mistake, 3.) Some sort of bizarre distro mistake, or 4.) Something you accidentally did.
If it's reproducible – e.g. with a new, virgin disk image -- I'd file a ticket.
Edit: From my node in Dallas, 192.168.1.1 is accessible but does not have an accessible DNS resolver.
Is there any problem with having a hostname as the full li123-456.members.linode.com vs li123-456 ?
Re: Resolver – it may be an artifact left over from when the image was built. However, if you are using DHCP the file should be overwritten. It's quite possible, however, that the file will be altered if you switch to static IP addresses. If you made this change, you should be manually setting your resolv.conf file anyhow.
Is there any problem with having the hostname as localhost? I don't plan on using that in production, just seemed to be the default today.
No, some distributions default to this if no hostname is set and if one cannot be obtained by DHCP. Your best bet is to set one manually as part of your system's configuration.