I gave my Vista system a similar configuration. The strange thing is, when
I "ping hostname" the first time, it resolves the name. Subsequent attempts
fail. If I flush the DNS cache (ipconfig /flushdns) it will work again -
once. Also, if I allow the cache to time out, it will also work again one
time. Another oddity is that when I display the DNS cache entries I do not
see the entry for "hostname".
Digging a little bit further, I sniffed the network traffic. On the initial
name resolution attempt, I see a standard query for an "A" record followed
by the server response, as I would expect. On subsequent attempts, I see 2
IPv4 LLMNR and 2IPv6 LLMNR probes, followed by 3 NBNS broadcast requests.
There is no DNS query.
Is this a bug? I would expect that after the initial attempt that Vista
would use the cached DNS info to perform the resolution, but this does not
appear to be the case. Any ideas?
Thanks,
-Jeff
| Vista Can't join domain because
single-label DNS ... It is not recommended to use
Single-label DNS name as domain name. In this case, you may
enable WINS so ... www.chicagotech.net/vista/vista&domain.htm |
"JDavis" <davis...@comcast.net> wrote in message news:TbGdnfD9NqfWhUfb...@comcast.com...
I suppose that the easiest thing to do would be to introduce a DNS domain
such as "local" into the mix. However, I am trying to understand this
behavior in Vista.
- Why doesn't Vista resolve the name from the local DNS cache after it is
initially resolved?
- Why doesn't the hostname appear in the "ipconfig /displaydns" output after
it is initially resolved? (see comments below)
- If the name isn't in the cache, why does flushing it have any effect
whatsoever?
If I disable client side DNS caching (net stop dnscache), the name resolves
every time.
It is interesting to note that in XP also the resolved name does not appear
in the cache. Rather, the XP client queries the name server every time.
For a name server, I am using dnsmasq on a DD-WRT router. It doesn't appear
that this is returning an authoritative response. I just notice that the
TTL on the inital nameserver response is 0, so that would explain why the
entry is not placed into the cache. It does NOT explain why resolution
fails after the initial attempt nor why flushing the cache makes the name
resolve again. It would appear to me that there is a flaw in the resolver
caching logic somewhere along the way.
-Jeff
"JDavis" <davis...@comcast.net> wrote in message news:ad-dnVR1tfiDDUbb...@comcast.com...