caching data

[Barry Margolin]

In article <9krsmk$jsu at pub3.rc.vix.com>,
Wittenberg, Chris <cwittenberg at ec2.edu> wrote:
>
>Since you mentioned this, I just have to ask:
>
>1. How would you 'confirm' or 'notice' that your cache is poisoned?

If you look something up and get an answer different from what the
authoritative servers for the domain are returning.

>2. What are the ramifications of dumping the cache (longer lookups while it
>rebuilds?).

Yes.

>3. Is there a recommended point at which you should run the 'rndc dumpdb' as
>a matter of course - or just when you think something is wrong?

As I mentioned in my other response, dumping the cache doesn't delete
anything.  You have to restart named, with "ndc exec" or "ndc restart".

>4. Is the cache dumped when you do a "kill -HUP pid"? Why is "rndc reload"
>preferred over kill (or is it), and does the "rndc reload" dump the cache?

No, none of these things touch the cache.  They tell named to reread its
configuration file, and update all the things that can be configured from
there (options can be changed, and zones can be added or deleted).  Since
the cache is not controlled by anything in the config file, it isn't
touched.

-- 
Barry Margolin, barmar at genuity.net
Genuity, Woburn, MA
*** DON'T SEND TECHNICAL QUESTIONS DIRECTLY TO ME, post them to newsgroups.
Please DON'T copy followups to me -- I'll assume it wasn't posted to the group.