Volker Kuhlmann wrote:
When I managed the DHL network for Europe, we had 2 old HP-UX boxes, on 10 mbit networking. They easily managed the load, but each had approximately the computing power of a 386SX25. We measured the uptime in years - in fact, they were so old, we didn't dare to reboot them!Thanks Jim for your insights. If I understand you correctly, the browser redoes the query every so often, and then the query to the ISP's name server fails for any of the reasons you gave.
Does your cache server consult the root zone servers directly, or does it talk to Xtra?
Paradise. I have cable, so using the ISP's servers makes more sense.
you will probably see no improvement ...
Hmm, that's because when the TTL expires at the ISP's cache, it would expire at the local cache too?
I find it a bit hard to believe that ISP's name servers are that overloaded. It would inconvenience an awful lot of their customers and occur with any browser. When konqueror tells me "host doesn't exist" after clicking on a link which stays on the same site, it does instantly. It is not waiting not-long-enough for a query to come back.
To test your hypothetis, what do I put into the forwarders setting in bind9 to bypass the ISP?
Thanks,
Volker
The amount of data that a DNS server needs to store is, by today's standards, trivial, and the size of the data transferred in the course of a dns loookup is so small, that I'd be looking for networking problems if a lookup fails when client and server are on the same subnet.
One thing that I have noticed is that these routers that offer everything... dns, dhcp, firewall, etc ( mine's an Alcatel Speedtouch 510 ), only offer a subset of the lookup services you expect. For example...
$> dig mx google.com
; <<>> DiG 9.2.4 <<>> mx google.com ;; global options: printcmd ;; Got answer: ;; ->>HEADER<<- opcode: QUERY, status: NOTIMP, id: 23272 ;; flags: qr ra; QUERY: 1, ANSWER: 0, AUTHORITY: 0, ADDITIONAL: 0
;; QUESTION SECTION: ;google.com. IN MX
;; Query time: 1 msec ;; SERVER: 10.0.0.138#53(10.0.0.138) ;; WHEN: Thu Feb 3 22:39:42 2005 ;; MSG SIZE rcvd: 28
NOTIMP = not implemented. This can cause you problems (:
( as opposed to
$> dig google.com
; <<>> DiG 9.2.4 <<>> google.com ;; global options: printcmd ;; Got answer: ;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 36502 ;; flags: qr rd ra; QUERY: 1, ANSWER: 3, AUTHORITY: 4, ADDITIONAL: 0
;; QUESTION SECTION: ;google.com. IN A
;; ANSWER SECTION: google.com. 300 IN A 216.239.37.99 google.com. 300 IN A 216.239.39.99 google.com. 300 IN A 216.239.57.99
;; AUTHORITY SECTION: google.com. 345600 IN NS ns4.google.com. google.com. 345600 IN NS ns1.google.com. google.com. 345600 IN NS ns2.google.com. google.com. 345600 IN NS ns3.google.com.
;; Query time: 256 msec ;; SERVER: 10.0.0.138#53(10.0.0.138) ;; WHEN: Thu Feb 3 22:41:17 2005 ;; MSG SIZE rcvd: 148
)
straight lookup is no problem. I just can't look up the mail exchanger record.
I'd build yourself a local dns server, plug it on the side of your workstation. It's only a small learning curve... actually pretty logical for a change. I've got the O'Reilly DNS and Bind book you can borrow if anyone's interested.
Cheers,
Steve.
