Reverse DNS question

Janos Dohanics web at
Tue Feb 18 23:26:43 UTC 2014

Hello List,

Could you please explain this odd behavior:

My Postfix logs show entries like this:

Feb 18 08:35:13 barrida postfix/smtpd[86649]: connect from unknown[]
Feb 18 08:35:13 barrida postfix/smtpd[86705]: connect from[]

This host is a source of legitimate messages, and sends a number of
messages every day. However, it seems that more often than not, Postfix
is unable to resolve the name for Postfix queries a
resolver (djbdns) which runs on the same machine.

I understand that DNS lookups can fail for reasons other than records
not existing. However, every time I check with host:

# host domain name pointer domain name pointer domain name pointer

or with dig:

# dig -x

; <<>> DiG 9.9.3-P2 <<>> -x
;; global options: +cmd
;; Got answer:
;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 32993
;; flags: qr rd ra; QUERY: 1, ANSWER: 3, AUTHORITY: 0, ADDITIONAL: 0


;; ANSWER SECTION: 38333 IN	PTR 38333 IN	PTR 38333 IN	PTR

;; Query time: 5 msec
;; WHEN: Tue Feb 18 17:41:23 EST 2014
;; MSG SIZE  rcvd: 130

I get replies as expected.

Of all the hosts which send e-mail regularly, this is the only one with
such odd behavior.

So, my questions are:

1. Other than network congestion, what might cause this recurring name
resolution failure?

2. If you look at the time stamps of the above 2 log entries: How is it
possible that precisely at the same time, name resolution BOTH does not
succeed AND does succeed? This "coinciding" time stamp isn't unique
either; I could show a number of other instances.

The system is FreeBSD 9.2-STABLE, postfix-2.10.2,1, djbdns-1.05.

Janos Dohanics

More information about the freebsd-questions mailing list