named still shuts down - help!

[Peter Rose]

Hi - 

I posted a couple of days ago that my named process was shutting itself
down inexplicably (well, inexplicable to me, anyway)

The problem seemed to go away pretty much by itself for 48 hours, so I
assumed maybe it was just some quirk of BIND 8.2.2 as suggested by Mark.

However, I have just had the process die 3 times in last couple of hours
- this is the output from the end of named.run at the time it went down
last:

req: nlookup(fine-sys.co.jp.cyberscreen.com) id 49814 type=1 class=1
req: found 'fine-sys.co.jp.cyberscreen.com' as 'cyberscreen.com'
(cname=0)
ns_req: answer -> [193.154.221.70].1086 fd=30 id=49814 size=113 rc=3
named shutting down
shutting down; dumping zones that need it
USAGE 979744041 979743670 CPU=0.14u/0s CHILDCPU=0u/0s
NSTATS 979744041 979743670 A=21 PTR=19
XSTATS 979744041 979743670 RR=49 RNXD=5 RFwdR=29 RDupR=0 RFail=0 RFErr=0
RErr=0

Unfortunately I am floundering as to what the stats may or may not tell
me.

Here's the background - I am running Bind 8.2.2 on RedHat 6.1 It's a
small server which is actually carrying the websites and the nameservers
on the same machine.(Yes, I know...). I have 8 IP's on this box, two of
which are given the names ns1.cyberscreen.com (212.87.82.28) and
ns2.cyberscreen.com (212.87.82.28). Now, for the last year (when I
started having to run my own dns) I have only ever started one named
process with one named.conf and everything has actually seemed to work
OK - BUT the syslog files for today's problem show the following on
restarting manually:
==================================================
Jan 17 15:07:03 server named[8439]: Lame server on 'fine-sys.co.jp' (in
'fine-sy
s.co.jp'?): [210.141.108.226].53 'dns1.dion.ne.jp'
Jan 17 15:07:21 server named[8439]: named shutting down
Jan 17 15:07:21 server named[8439]: USAGE 979744041 979743670
CPU=0.14u/0s CHILD
CPU=0u/0s
Jan 17 15:07:21 server named[8439]: NSTATS 979744041 979743670 A=21
PTR=19
Jan 17 15:07:21 server named[8439]: XSTATS 979744041 979743670 RR=49
RNXD=5 RFwd
R=29 RDupR=0 RFail=0 RFErr=0 RErr=0 RAXFR=0 RLame=1 ROpts=0 SSysQ=16
SAns=29 SFw
dQ=13 SDupQ=10 SErr=0 RQ=40 RIQ=0 RFwdQ=0 RDupQ=0 RTCP=0 SFwdR=29
SFail=0 SFErr=
0 SNaAns=15 SNXD=10
Jan 17 15:11:21 server named[8527]: starting.  named 8.2.2-P7 Fri Nov 10
11:12:5
3 EST 2000 ^Iprospector at porky.devel.redhat.com:/usr/src/bs/BUILD/bind-
8.2.2_P7/s
rc/bin/named
Jan 17 15:11:21 server named[8527]: Zone "cyberscreen.com" (file
cyberscreentest
1.hosts): No default TTL set using SOA minimum instead
Jan 17 15:11:21 server named[8527]: master zone "cyberscreen.com" (IN)
loaded (s
erial 2000110601)

[it then loads all the zones OK - yes, I really must fix the $TTL
sometime:-(]

Then I get this, which I have not seen before:


Jan 17 15:11:21 server named[8527]: bind(dfd=20, [127.0.0.1].53):
Address alread
y in use
Jan 17 15:11:21 server named[8527]: deleting interface [127.0.0.1].53
Jan 17 15:11:21 server named[8527]: bind(dfd=20, [212.87.82.24].53):
Address alr
eady in use
Jan 17 15:11:21 server named[8527]: deleting interface [212.87.82.24].53

and so on for all my IP's - THEN it logs:

Jan 17 15:11:21 server named[8527]: not listening on any interfaces
Jan 17 15:11:21 server named[8527]: Forwarding source address is
[0.0.0.0].2896
Jan 17 15:11:21 server named[8528]: Ready to answer queries.
Jan 17 15:11:26 server named[8528]: Sent NOTIFY for "abechi.co.uk IN
SOA" (abech
i.co.uk); 1 NS, 1 A

and so on for all my zones.