Bug #7484

facter fqdn is empty when domainnames aren't containing any dot (ie tld)

Added by Edouard Bourguignon about 1 year ago. Updated 8 days ago.

Status:Closed Start date:05/11/2011
Priority:Normal Due date:
Assignee:Hailee Kenney % Done:

0%
Category:library
Target version:2.0.0
Keywords:tld fqdn hostname dot intern Affected Facter version:
Branch:https://github.com/puppetlabs/facter/pull/199
Votes: 1

Description

As far as I understand the following regexp used in domain.rb to find out the domainname:

    domain = Facter::Util::Resolution.exec('dnsdomainname')
    next domain if domain =~ /.+\..+/

domainname must contain a dot, which is not completely compliant with RFCs[¹] and prevent using facter on tld servers.

How to reproduce: On a server with fqdn “ns01.tld”, this is the result of some commands:

  • hostname => ns01 (OK)
  • hostname -f => ns01.tld (OK)
  • facter hostname => ns01 (OK)
  • facter fqdn => (bad)
  • dnsdomainname => tld (OK) but dnsdomainname command had to be installed

facter fqdn should return ns01.tld and should be (very important) consistent with “hostname -f”.

     "old-simple-name.ARPA".  That is, the all the hosts are in a
     domain called "ARPA"

Related issues

related to Facter - Bug #3909: Strip trailing dots from domain fact Closed 05/28/2010
duplicated by Facter - Bug #11229: On a clean, minimal install of Debian 6, `facter domain` ... Duplicate 12/06/2011
duplicated by Facter - Bug #3356: fact domain empty if domainname doesn't contain a dot (.) Duplicate 03/09/2010

History

#1
Updated by Daniel Pittman about 1 year ago

  • Status changed from Unreviewed to Accepted

Thanks for this report, and the detailed problem analysis.

facter fqdn should return ns01.tld and should be (very important) consistent with “hostname -f”.

Can you expand on why that, specifically, is so important? That way we can be sure we understand your use-case and reasoning.

#2
Updated by Edouard Bourguignon about 1 year ago

It’s important because other unices or applications/daemons are already using the hostname command to determine the short and full hostname (fqdn). So if this values are not the same than the facter hostname/fqdn, it could be problematic. We don’t want to use any command from dns client like dnsdomainname or modifying /etc/resolv.conf (they are simply not available).

#3
Updated by James Turnbull 9 months ago

  • Category set to library

#4
Updated by Ken Barber 6 months ago

  • Target version set to 1.6.x

#5
Updated by Jeff Weiss 20 days ago

  • Keywords changed from tld fqdn hostname dot to tld fqdn hostname dot intern

#6
Updated by Jeff Weiss 16 days ago

  • Assignee set to Hailee Kenney

#7
Updated by Jeff Weiss 16 days ago

it looks like this would pull the tld from an /etc/resolv.conf if the domain or search fields existed in it. Granted this is probably not the behaviour we want to encourage.

#8
Updated by Jeff Weiss 16 days ago

It’s rumoured that on Solaris hostname -f will set the hostname to -f. Hailee and I will look into this because it could adversely affect making the domain fact always match hostname -f.

#9
Updated by Jeff Weiss 15 days ago

The Solaris hostname man page indicates Solaris run with hostname -f will indeed alter the system’s hostname to -f.

#10
Updated by Hailee Kenney 15 days ago

  • Status changed from Accepted to In Topic Branch Pending Review
  • Branch set to https://github.com/puppetlabs/facter/pull/198

#11
Updated by Hailee Kenney 14 days ago

  • Status changed from In Topic Branch Pending Review to Code Insufficient
  • Target version changed from 1.6.x to 2.0.0
  • Branch deleted (https://github.com/puppetlabs/facter/pull/198)

Using hostname -f is a potentially breaking change, so it should be targeted at Facter 2.0 instead of 1.6.x.

Also, HP-UX doesn’t handle hostname -f, so changing it to opt-in instead of opt-out.

#12
Updated by Hailee Kenney 14 days ago

  • Status changed from Code Insufficient to In Topic Branch Pending Review
  • Branch set to https://github.com/puppetlabs/facter/pull/199

#13
Updated by Hailee Kenney 14 days ago

  • Status changed from In Topic Branch Pending Review to Merged - Pending Release

Documentation added to pre-docs.

#14
Updated by Hailee Kenney 14 days ago

  • Status changed from Merged - Pending Release to Code Insufficient

#15
Updated by Hailee Kenney 14 days ago

  • Status changed from Code Insufficient to In Topic Branch Pending Review

#16
Updated by Daniel Pittman 11 days ago

  • Status changed from In Topic Branch Pending Review to Merged - Pending Release

We believe that this has been fixed, and should be resolved in the release of Facter 2.0.0

#17
Updated by Matthaus Litteken 8 days ago

  • Status changed from Merged - Pending Release to Closed

released in Facter 2.0.0rc1

Also available in: Atom PDF