named-xfer causes 100% CPU Utilization

[Danny Mayer]

At 11:07 AM 9/28/01, Kevin Vaughn wrote:
> > There was a bug in the NT code that caused this to happen. This is fixed
> > in 8.2.5 which will be available shortly. The real problem is with the
>primary.
> > Look in it's event log for errors.  Try upping the pagefile size and
>rebooting the
> > primary.  It may very well be running out of memory during the zone
>transfer.
>
>The only error I have on the primary server is:  "named.conf:8: syntax error
>near allow-update"
>
>This is the code:
>
>options {
>  directory "c:\var\named";
>  allow-update {10.5.90.4;};
>  };
>
>Above is the only error I am getting on either server. 

allow-update is not valid for the options statement.  This is specific to a zone.
Only zones should allow dynamic updates. Did you mean allow-transfer?
That would be a valid statement for the options.

>  The pagefile for
>this system is static at 250MB.  I have rebooted several times, but the
>slave server is still behaving badly.

This is quite low. For the primary, I'd try 500MB, then reboot and then
see if the transfer works.  Are there other applications running on the
primary?


> > I assume you really mean hosts in the domain.
>
>Sorry... That's the newbie coming out ;)
>
>We have around 1,000 hosts.  We have around 120 reverse arpa files.  We have
>1 zone.
>
> > No, but the problem has been fixed.
>
>I realized that the 8.2.5 release candidate was out.  Has anyone produced an
>NT port of this version yet?  If not, then my immediate problem hasn't been
>fixed (unless I want to build a Linux box, which isn't out of the question).

I think you meant built the NT Binaries. The port already exists in the tarball.
Building is a matter of running the build process.

I have been building binaries all along, just a few unrelated issues have
prevented it from being released.

         Danny