ndc restart probs?

[Kevin Darcy]

Just out of curiosity, why are you doing a "ndc restart" as part of an
automated maintenance system? "ndc reconfig" would seem to be a better choice
to make "named" discover added or deleted zone definitions, and "ndc reload
<zone>" would seem to be a better choice for making it reload a specific zone.
Are you really changing *global* configuration values that often?


- Kevin
Riskable wrote:

> Riskable's Stationary
>
>       HOME
>
>       EMAIL ME
>
>       PAGE ME
>
>       NEWS
>
>       I am running two master servers (NS1 and NS2).  Each one is host to
> close to 400,000 domain names.  Both are running the latest version of BIND
> (release, not beta).
>
>       For some reason, the "ndc restart" command stopped working on both of
> them.  It just sort of hangs and the server stops responding to DNS
> requests.
>
>       The fix is to simply kill named first, then use "ndc start".  However,
> I was just wondering if anyone else had this problem?  Sometimes even "ndc
> start" causes the same problem but if I just wait a little while it seems to
> work fine again (say, five minutes?).
>
>       Oh, and for some odd reason the "ndc start" command on NS1 reports
> back "named has not started yet(?)" but everything seems to work fine.  Just
> a quirk IMHO, but it might have give some clues to my problem?
>
>       Normally this problem wouldn't be a big deal but it is SERIOUSLY
> annoying when trying to automate the updating of the DNS servers (which is
> supposed to happen every 20 minutes).
>
>       named, at it's highest point, uses like 15% of the CPU and 50% of the
> memory so I don't think the machine is simply overloaded.  I just ordered an
> extra 128 megs.
>
>       -Riskable
>       http://riskable.youknowwhat.com