On Saturday, October 26, 2002 4:40 PM, Danny Mayer <mayer@xxxxxxx> wrote:
>
>At 10:43 AM 10/24/02, Bob McGregor wrote:
>>below
>>
>>On Wednesday, October 23, 2002 6:44 PM, Danny Mayer <mayer@xxxxxxx> =
wrote:
>> >
>> >At 10:49 AM 10/23/02, Bob McGregor wrote:
>> >> >C:\WINNT\system32\DNS
>> >> >>
>> >> >>I'll do the dig command when it fails to respond again.
>> >> >>
>> >> >>thanks danny, bob
>> >> >
>> >> >Okay.
>> >> >
>> >> >Danny
>> >> >
>> >> >
>> >> >>On Friday, October 18, 2002 7:15 AM, Danny Mayer <mayer@xxxxxxx> =
wrote:
>> >> >> >
>> >> >> >At 11:13 AM 10/15/02, Bob McGregor wrote:
>> >> >> >
>> >> >> >>This morning it stopped resolving before I did the stop/restart =
the =3D
>> >> >> >>netstat -n listed fewer IP connections to the server(5 external =
IP's
>> >> none =3D
>> >> >> >>to dns ports). Is there something specific I'm looking for?=
=3D20
>> >> >> >>
>> >> >> >>Also the rndc status liksted nothing different than when it was
>> >> >> resolving, =3D
>> >> >> >>same speed no different information.
>> >> >> >
>> >> >> >I probably should have asked you to do a TCP query using dig =
like:
>> >> >> >dig +tcp SOA domain @server
>> >> >> >
>> >> >> >and see if it responds. I suspect that it's the UDP part that's =
no=20
>> longer
>> >> >> >listening.
>> >> >> >What's the output of netstat?
>> >> >> >
>> >> >> >Danny
>> >> >> >
>> >>
>> >>The server stopped responding again this morning 8:40am MST. here are=
>> >>some results from the commands. It would resolve our internal =
numbers but
>> >>would not resolve external addresses. After the stop/restart, it =
resolved
>> >>again.
>> >>all the following command were performed before the stop/start...
>> >>bob
>> >
>> >This reply indicates that the problem really is that it is not able to=20
>> retrieve
>> >EXTERNAL DNS records since you are able to get responses for your
>> >own authorative data. Your original post seemed to indicate that it
>> >wasn't responding to anything at all.
>> >
>> >This all sounds more like a connectivity problem. If you do a dig for
>> >www.hp.com using +tcp do you get an answer?
>> >
>> >dig +tcp www.hp.com A
>> >
>> >Danny
>> >
>>yeah in a later post yesterday I included the dig info you mention...=20
>>sorry I forgot it in the 1st one.
>>
>>Also what I figured out is that it responds from the keyboard of the NT=20
>>box as it uses 127.0.0.1 by default but when I dig forcing it to use the=20
>>public address it fails and it fails from all other clients...
>>
>>however when I add the +tcp, it works with the public address reference =
as=20
>>well.... I think you were right in a prior post that the UDP port stops=20
>>responding which is causing the problem...
>>
>>when I did the netstat -an it listed the following for port 53 for the=20
>>public IP of the server:
>> TCP 207.196.133.10:53 0.0.0.0:0 LISTENING
>> TCP 207.196.133.10:53 0.0.0.0:0 LISTENING
>> UDP 207.196.133.10:53 *:*
>>this is the same info as when the server is working correctly....
>>
>>bob
>
>Are you running NT 4.0 or W2K? There is a problem in W2K which causes
>it to stop responding on a UDP socket. This is due to a Microsoft bug
>that they introduced in W2K in the protocol. There is a workaround for it
>in SP2, but it also requires code changes to BIND, instead of their just =
fixing
>the underlying bug. There is a fix for this in BIND 9.3.0.
>
>Danny
>
thanks Danny,
I'm running on 4.0 sp6. Where do I get 9.3.0 for NT? Do you think it =
will
fix my problem?
thanks, bob
|
