Thank you for your so helpful informations (from Hungary?)
On 01/25/02(21:58) you "Ferenc Toth" <firstname.lastname@example.org> wrote
|>>One more thing. I've noticed that the RecvFrom errors are logged right
|>>after some busy time. So if delegate services multiple queries in a
|>>very short time, a RecvFrom error will follow. Don't ask me why are
I'm reminded that I did see the infinite loop was caused on such
contiguous requests from specific clients. But I could not reproduce
it because I could not what was specific in those clients.
|>>there multiple queries for the same domain name. (192.168.99.5 is an
|>>internal linux machine running bind9.2 servicing internal DNS queries
|>>and forwarding all other requests to delegate).
|Looks like, I've got it. We had some misconfigured services asking
|stupid invalid DNS queries from the ISP, also the user activity has
|risen since we upgraded from 128kbit to 256kbit recently. I've fixed
|the misconfigured services and added "DNSCONF=para:8" to the dns proxy.
|Tested it with a few hundred almost simultaneous requests and so far it
|looks to be ok! :)
Could you tell me how the query was invalid?
I'll try to modify DeleGate to be more robust on any invalid queries.
@ @ Yutaka Sato <email@example.com> http://www.delegate.org/y.sato/
( - ) National Institute of Advanced Industrial Science and Technology (AIST)
_< >_ 1-1-4 Umezono, Tsukuba, Ibaraki, 305-8568 Japan