At 01:16 25/09/2008, JINMEI Tatuya / =?ISO-2022-JP?B?GyRCP0BMQEMjOkgbKEI=?= wrote: >At Mon, 22 Sep 2008 07:24:02 +0100, >Jan Arild Lindstrøm <[EMAIL PROTECTED]> wrote: > >> >Second, limiting max-clients-per-query doesn't help reduce the number >> >of recursive clients if the same query is sent from different IP >> >addresses. >> >> Auch! Is that really correct? Should it not then be called >> "max-queries-per-client" and >> not "max-clients-per-query"? > >Oops, I was wrong. I was confused about the case where a single (or >multiple) client keeps sending a high volume of different bogus >queries (for which max-clients-per-query doesn't help). > >I now see the problem. It's really strange to have more than 10,000 >recursive clients for the same query while max-clients-per-query is >100. I have no specific idea about how this could happen, but I'd >suspect this may be a thread-related bug. Is it possible to rebuild >named without threads and see if the same problem happens?
I do not know if our T2000 will manage to reply fast enough, or if it fast will get saturated if it uses only one thread for alle the traffic. I will rebuild and test a little in parrallell on one of them. Regards Jan Arild Lindstrom
