Slave won't update with BIND 8.2.3 on NT

[Danny Mayer]

At 12:57 PM 3/7/01, Tim Madden wrote:

>On 6 Mar 2001, at 23:06, Danny Mayer wrote:
>
> > At 09:19 PM 3/6/01, Tim Madden wrote:
> > 
> > >I've just set up my first slave server, but named-xfer won't complete
> > >when called by named.  I can manually run named-xfer and it will pull in
> > >the zone.  I didn't use the default directory structure, so I wonder if
> > >that's part of the problem?  When named calls named-xfer, I can watch it
> > >start and stop in Task Manager.
> > 
> > Not related.  It really doesn't care as long as it can read and write to the directory.
>
>That's what I expected, but it's nice to hear someone else say it.
>  
> > >After each failure, named-xfer strands the following files in my C:\BIND
> > >directory:
> > >
> > >db.aspenresearchgroup.com.dom.000044
> > >db.aspenresearchgroup.com.dom.ixfr.tmp.000044
> > >xfer.ddt.000044
> > >
> > >where 000044 changes, I assume, with the PID for each failed named-xfer.
> > >
> > >The first two files are empty, but the xfer.ddt.xxxx file contains:
> > >domain `aspenresearchgroup.com'; file `db.aspenresearchgroup.com.dom';
> > >serial 0
> > >domain `aspenresearchgroup.com'; ixfrfile
> > >`db.aspenresearchgroup.com.dom.ixfr.tmp'; serial 0
> > >tsigfile `aspenresearchgroup.com.362'
> >
> >          These are standard temporary files that I created by the
> > transfer. They are left behind because of the transfer failure. 
>
>Again, what I expected.  What about the reference to ixfr?  I know 
>there's an axfr and and ixfr, but I don't know the difference between 
>them.  Could the manual copy be doing and axfr and the automatic 
>transfer is failing because it's doing an ixfr?

         ixfr is an incremental transfer and has to be explicitly enabled otherwise
it will do just a axfr.  Don't worry about the name, that happens to be just the
way the code set up the file name. The file is supposed to disappear once the
zone is loaded.

>  
> > >I've successfully used the command line:
> > >
> > >c:\bind> bin\named-xfer -z aspenresearchgroup.com -f
> > >db.aspenresearchgroup.com.dom -s 0 192.168.50.16
> > >
> > >I've also successfully transferred the zone with -s 2001030602 (the
> > >master's ser# is 2001030604)
>
>One thing I didn't notice until after I sent my message is that the 
>manual transfer copies the file, but doesn't load the zone.  Should 
>it load the zone?

         Manual transfers won't load the zone. The server knows nothing about the
zone. If you do manually transfer the zone, you can copy it to the location that
you indicated for the zone in named.conf and then restart the server.

> >          These transfers don't check the zone contents. What does
> > the application event log tell you.  You should see record errors
> > logged there. 
>
>Nothing, really.  A few informational "interogated" messages and 
>one warning that the directory is "world-writable". However, it does 
>say I'm loading BIND 8.2.3-T6B-NT Nov 5, 2000.  I've just re-
>downloaded and rechecked all my file dates.  All my executables 
>are dated 1/27/01, the same as what is in the newly downloaded 
>.zip file, so I can only assume I'm on the released version?

         If this is what it says in the event log, then that's what you are running.
Try reinstalling again.  You don't need to stop the service, BINDInstall will do
that for you. The version should say BIND 8.2.3-REL.


> >          Since your master for the domain is in a private subnet I can't check, but
> > I bet you have a CNAME at the root of the zone.  If so, you need to get rid of it
> > as it's illegal and will prevent the slave from loading the zone.
>
>Hmm...a CNAME at the root of the zone?  That I don't understand.  
>  With the risk of being presumptuous, I've inclued the zone file from 
>the master at the bottom.

         I looked at this.  I can't see anything obviously wrong.  I may just load it
up on my own system and see if I get any errors.

>Also, today I've noticed an error on my private, behind the firewall 
>name servers (the ones we've been talking about).  They both 
>report 
>
>23-Feb-2001 18:19:54.000 default: info: sysquery: findns error 
>(NXDOMAIN) on ns2.aspenres.com?
>
>and 
>
>07-Mar-2001 08:05:39.000 default: info: sysquery: findns error 
>(NXDOMAIN) on ns1.aspenres.com?  
>
>ns1 and ns2 are my public name servers outside my firewall.  I'm 
>into http://www.acmebw.com/askmrdns/bind-messages.htm 
>looking for an clue now, but any suggestions welcome.  I must 
>have something misconfigured, but I'm not sure what.

         You need to forward all queries to a DNS that can make queries on the outside.
  Use the forwarders and forward statements for this.

         Danny