How to know the reason of SERVFAIL

[Barry Margolin]

In article <df4ucm$m3r$1 at sf1.isc.org>,
 "Watanabe, Hiroshi (HP-Partner)" <hiroshi.watanabe at hp.com> wrote:

> I'm using BIND9.2.3 on True64 Unix as cache server.
> When I query some domain, status SERVFAIL is returned.
> 
> How can I know the details of reasons for SERVFAIL.
> For example,  "No NSs with CNAMEs",  "Lame nameservers" ... so on.
> 
> Is there any logging method for this?

There's no single technique, because it depends on where the SERVFAIL is 
coming from.  If it's coming from the authoritative server, you will 
probably be able to tell from the log messages that were generated at 
the time it attempted to load the zone file.  But if it's coming from a 
caching server, there may or may not be relevant log messages.  For 
instance, it will report SERVFAIL if none of the authoritative servers 
respond, but there's nothing logged in this case.  It may also report 
this error if the server is lame, but you won't get log messages if you 
disable lame-server logging (as many sites do, since these are so common 
and there's nothing you can do about it).

-- 
Barry Margolin, barmar at alum.mit.edu
Arlington, MA
*** PLEASE post questions in newsgroups, not directly to me ***