RE: Connection failed named

[Mike Robertson]

if failed host 127.0.0.1 port 53 type tcp then restart



-----Original Message-----
From: address@hidden
[mailto:address@hidden On Behalf
Of Iann Gorrill
Sent: Monday, August 21, 2006 10:27 AM
To: address@hidden
Subject: Connection failed named

Hello everyone,

Probably a simple answer somewhere to this question, but I just can't 
find it.  Using the examples on the site, I setup monit to monitor my 
named installation:

  check process named with pidfile /var/run/named/named.pid
    start program = "/etc/init.d/named start"
    stop program = "/etc/init.d/named stop"
    if failed port 53 type tcp protocol dns then alert
    if failed port 53 type udp protocol dns then alert
    if 5 restarts within 5 cycles then timeout

However, when I start monit, I get an alert that the test failed.  When 
running monit in debug mode, I see this:

'named' zombie check passed [status_flag=0000]
'named' succeeded connecting to INET[localhost:53] via UDP
DNS: no answer records returned
'named' failed protocol test [DNS] at INET[localhost:53] via UDP
Connection failed notification is sent to address@hidden
'named' succeeded connecting to INET[localhost:53] via TCP
DNS: no answer records returned
'named' failed protocol test [DNS] at INET[localhost:53] via TCP

It would appear to correctly connect to the port, but it isn't liking 
the answer it gets to its question.  So, I turned on querylogging and 
increased the trace level in bind, and here's what I saw in 
/var/log/messages:

Aug 21 09:00:03 webgate named[21157]: client 127.0.0.1#43170: query: . IN NS

Well, that's the only one of any interest to my cause.  A tcpdump didn't 
seem to show much else either.  If I set the host to an external 
nameserver in the monit.conf, it works, so I'm not sure what other 
direction to go here.
-- 
Iann


--
To unsubscribe:
http://lists.nongnu.org/mailman/listinfo/monit-general