HPC Public Node Naming Convention

BC Project: FY11-02
Date of Policy: 22 Aug 2011
Last Updated: 11 Jun 2019 (see Revision Log)

This policy provides the user with a consistent way to access user-accessible systems at all participating HPCMP centers by using the name of the system and a uniform node naming convention.

Round Robin Implementation

Users will be able to log into a system by specifying the system name and domain. The specific login node on which the user will be placed is determined by the system in a round-robin fashion. The following example session using the ssh command illustrates this:

ssh excalibur.domain

[ ... message of the day and other messages ... ]

excalibur01>

Login Node Naming

The naming convention for login nodes will be the following:

systemnameNN (where NN ranges from 01 to number of login nodes on the system)

The following example session illustrates accessing login node 03 using the ssh command:

ssh excalibur03.domain

[ ... message of the day and other messages ... ]

excalibur03>

Login Node Availability

All efforts should be made by all Centers to exclude non-functional login nodes from the DNS round-robin. Any login node undergoing maintenance or extended fail-fix investigation, expected to last longer than 48 hours, should be removed from the DNS round-robin, thus ensuring a successful connection to the login nodes.


Revision Log
Date Revision
11 Jun 2019Added a new section titled "Login Node Availability"
15 Mar 2018BC Team Audit - Remove specific system domain. Wording and format simplifications
09 Jun 2016BC Team Audit
10 Apr 2014BC Team Audit
18 Jul 2013Reworded for clarity
02 Apr 2012BC Team Audit