On Nov 28, 2007 5:54 PM, Brendan Oakley <[EMAIL PROTECTED]> wrote: > > I am running this patch on production servers with very good results. > In fact it is a requirement in our environment because these are > heavily loaded servers and the recursor will bog down very badly and > even cause the server to falter when a single forwarder becomes > unavailable, even for the few seconds it can take for the forwarder to > reload a large zone. Having a second one available saves it, and it > never misses a beat.
Because I said this, I should revise it for the sake of accuracy. I still think the patch is working just fine, but it did not exactly solve my problem. I think it does better than it did, but I still a good percentage of the time, when the forwarder's zone is copied I get hundreds of log entries like: Recursive query for remote 10.2.14.105 with internal id 682 was not answered by backend within timeout, reusing id Then the authoritative server becomes very sluggish and fails to answer queries until the pdns and recursor daemons are restarted. I have the authoritative PowerDNS server, the recursor, and rbldnsd running on the same server, which could be my mistake. The recursor and rbldnsd are listening on different ports on localhost. I've tried many options without much luck. I keep thinking I've solved it, but the next day it fails again. Anyway I just wanted to be accurate. I still think the multiple forwarder patch is working as it should. Thanks. Brendan _______________________________________________ Pdns-users mailing list [email protected] http://mailman.powerdns.com/mailman/listinfo/pdns-users
