On 5/27/20 11:37 PM, Thomas Mieslinger via Pdns-users wrote: > Hi Mike, > > > Anyone want to (gently) shoot me down....? > > not really. > > Of course, the correct fix would be to fix the authoritative setup. > > You could do some research on your end to see how many requests your > servers are doing towards the service provider. Maybe you run in some > sort of rate limiting. > > On the other hand I have failed to remove my blacklisted recursor source > ips from national-lottery.co.uk/nsX.camelotinteractive.com (speaking of > the recursors used by gmx.net/web.de/mail.com mail system) to be able to > resolve the MX records for national-lottery.co.uk. > > My fix was to add a forward to cloudflare/google dns for this (in my > opinion bad behaving) domain to empty up the mailqueues and stop > customer complaints.
The site in particular that touched off my note, was 'wildblue.net', and for the same reasons - backed up mail queue, which got me to thinking this is an occasional problem and if we just had some kind of cache of the cache we could reduce this kind of issue and provide better service to end users as a whole. As Klaus Darilion pointed out, and to my surprise, bind actually implements this feature. In my case, my mailserver setup uses a local instance of dnsmasq which forwards to my normal resolvers, and so I was able to drop a quick config file in with 'server=/wildblue.net/8.8.8.8' and my queue cleared up pretty quick. I dont have time to chase down the admins for that network and explain the perils of single subnet auth name servers and so forth, just wanted a solution that likely would address > %80 of these cases. Thanks for the input. Mike- _______________________________________________ Pdns-users mailing list Pdns-users@mailman.powerdns.com https://mailman.powerdns.com/mailman/listinfo/pdns-users
