[CLUE-Tech] bind problem revisited

[Jeremiah Stanley]

> Stopping named: [  OK  ]
> Starting named: [  OK  ]
> 
> That makes me think it's back to normal, but it turns out that a named process is still left running, even after I restart it the second time.  I'm not sure why I'm getting this, but I do know that it is now the cause of named to not completely get out of memory, thus the reason that the old zone serial numbers are being transferred rather than my updated ones. 

Ok, correct me if I'm wrong but a restart will first stop the daemon and
then start it back up again. Do you mean something different like that
the process never died the first time?

Try this:

service named stop

you should get:

Stopping named: [  OK  ]

if it is running, it will fail if it is not running.

Now do:

service named status

On my machine I got:

rndc: connect failed: connection refused

and when it is running I get:

number of zones: 6
debug level: 0
xfers running: 0
xfers deferred: 0
soa queries in progress: 0
query logging is OFF
server is up and running

The next step that I would take is to stop named and then try doing a
'ps -A | grep named' and see what you get. If you stopped it then you
should have no named processes. If you still have one you may have
jumped a PID file somewhere in the past. Go ahead and kill that process
if it is running when you don't want it to be. If for some reason it is
running when you start the daemon, it will whine about not being able to
commandeer the socket or something to that affect.

If named is crashing alot I would check and see if there is a patched
version out there (I'm sure you already did this, right?). How many
queries are you taking per hour? A ton, a handful? If you are not taking
that many I would start blaming RAM defects and random acts of god.

-- 
JStanley <miah at miah.org>
http://www.slavewage.com/