Naming conventions

[jinxer]

How does your naming convention for devices look like? I always try to use the following syntax for instance:

<location>-<function>-<type>-#

Example;

sfu-c-sw-01

Where 'c' stands for core 'sw' for switch and so forth.

sit-mgmt-fw-01

Etc


Sent from my iPhone using Tapatalk

[icecream-guy]

don't have one,

name it what you want, edit /etc/hosts on the jumpbox and yer good...

[deanwebb]

We have a standard naming convention... and about three or four other standard naming conventions, so it depends on the project staff how a new box gets named. Somewhere in there, though, is the location and function of the box, if it's a network device.

Otherwise, we have names that are... unique...

[javentre]

name it what you want, edit /etc/hosts on the jumpbox and yer good...

We use DNS.  Manually configuring thousands of entries in an /etc/host file AND in DNS is a good way for one of the two to become askew, resulting in more issues. 

We also want it in DNS so traceroute (from any host) shows useful data.

[Seittit]

Geographical site code -> os type -> server function -> number

example, a file and print server in Japan would look like:
JAPMSNETFNP011
Japan
Microsoft
Network file and print
Number


Sent from my iPhone using Tapatalk

[Network2501]

<device>.<pop>

Based on the device you'll know what function and hardware it is. Here's lucky hop number 7 from an mtr.

Code Select Expand

7. ae0.cr1.mel4.on.ii.net                                                                             0.0%     9   51.0  53.3  50.7  68.5   5.7

This works well for networking kit but once you start dealing with systems or virtual machines it doesn't work so well. That said I don't think there needs to be one naming convention that's applied to everything.

[Bit_Jockey]

Previous administrative staff saw fit to name production boxes after lord of the rings characters, they even configured motd with about 2 paragraphs of story about each character.

Having never seen any of the movies, I believe they are all a bunch of nerds.

[deanwebb]

Having never seen any of the movies,

I believe they are all a bunch of nerds.

Totally. Of course, working on routers and stuff is kinda sorta also nerdy.

[jinxer]

name it what you want, edit /etc/hosts on the jumpbox and yer good...

We use DNS.  Manually configuring thousands of entries in an /etc/host file AND in DNS is a good way for one of the two to become askew, resulting in more issues. 

We also want it in DNS so traceroute (from any host) shows useful data.

Yea we too try to keep it to DNS instead of host files.. for much of the same reason.

[SimonV]

<three letter sitecode>-<hall or building number>-<type of device>-<number>

For example: LON-H03-AS01 or LON-DC-FW-01

This is my preference but I'm curently at a gig which has no real standard. Also when you're consulting you usually have to go with what the guys before you decided.

[icecream-guy]

name it what you want, edit /etc/hosts on the jumpbox and yer good...

We use DNS.  Manually configuring thousands of entries in an /etc/host file AND in DNS is a good way for one of the two to become askew, resulting in more issues. 

We also want it in DNS so traceroute (from any host) shows useful data.

Yea we too try to keep it to DNS instead of host files.. for much of the same reason.

We keep it out of DNS for security reasons,  one bad hack to your DNS server and your whole network topology is pretty much known.

[javentre]

The same can be said for keeping it in an /etc/hosts file, or keeping it in a visio diagrams, or in an .xls sheet too, right?

You can gather a ton of data with just a few trace routes too.

You can keep it in your internal DNS servers, while excluding it from your external ones.

[srg]

Of course this is kept on internal  DNS in own zones.

[that1guy15]

For enterprise or remote sites:
State-siteID-closet-function01 02 03 etc.

For the DC replace closet with rack number. We also place every IP in DNS for clean and simple traceroute and such.

[javentre]

We also place every IP in DNS for clean and simple traceroute and such.

We put the Loopback0 (mgmt) interface both the forward and reverse zones, but the other interface IPs only get reverse entries for traceroute resolution.