pcp
[Top] [All Lists]

Re: pmhostname failing - urgent

To: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Subject: Re: pmhostname failing - urgent
From: "Frank Ch. Eigler" <fche@xxxxxxxxxx>
Date: Fri, 15 Nov 2013 18:00:44 -0500
Cc: pcp@xxxxxxxxxxx
Delivered-to: pcp@xxxxxxxxxxx
In-reply-to: <52868ED9.90209@xxxxxxxxxxxxxxxx>
References: <000201cee1ed$6a1171f0$3e3455d0$@internode.on.net> <y0m7gc91x0f.fsf@xxxxxxxx> <5286702D.1000101@xxxxxxxxxxxxxxxx> <20131115193341.GD8993@xxxxxxxxxx> <5286804D.5010107@xxxxxxxxxxxxxxxx> <20131115203919.GE8993@xxxxxxxxxx> <52868ED9.90209@xxxxxxxxxxxxxxxx>
User-agent: Mutt/1.4.2.2i
Hi, Ken -

> >>fqdn=Unknown Host host=gwdb02.prod.mel.au.mydomain.com.au
> >>and the "Unknown Host" causes the awk failure at the end
> >
> >I don't understand how that message could be coming from a current pcp
> >installation, as we don't print "fqdn=" anywhere.
> 
> It is not the latest ... it is 3.8.0 (zero)

Ah, rereading your original message, I see I brought up the newer
versions, not you.  Sorry about my misunderstanding.

One of the several reasons we are gradually ditching pmhostname is
because of its assumptions about DNS/hostnames.  One of these is that
if you give it a host that can be resolved (to an A/AAAA record), it
will then proceed to feed that IP address back into DNS to map to a
PTR record, and give you that name.  It's assuming that forward &
backward DNS both will work and produce a unique & sensible answer,
and they will live happily ever after.

I might try tcpdump'ing port 53 to see the DNS traffic actually
emanating from pmhostname.  See if /etc/nsswitch.conf might have lost
the "dns" part.  See if your local DNS hierarchy (/etc/resolv.conf
etc.) has lost the custom PTR zones.

- FChE

<Prev in Thread] Current Thread [Next in Thread>