Klaus, Sorry for replying late.
> On Jan 17, 2018, at 7:32 AM, Klaus Darilion <[email protected]> > wrote: > > > > Am 17.01.2018 um 11:31 schrieb Steve Zeng: >> I came across this post and was concerned if the high number of NOTIFY/AXFR >> overloaded PowerDNS, given that we have ~6,000 zones and ~100 BIND slaves. >> Do you know if there is a built-in limit on the AXFR volume? > > Probably not the number of zones but the number of changes in this zone > is more relevant. > > Nevertheless, using that number of zones and name servers I would also > use a database replication (which implies that you have to replace Bind > with PowerDNS on your anycast nodes). I can recommend PostgreSQL with > Slony (works great here with 40 slaves and nearly 1mio zones) > > I wonder, in your current setup, are the slaves all in sync with the > master? That was our concern when we decided to use PowerDNS with > DB-replciation. > most of the slaves can keep in sync except a few frequently updated zones which is also the reason I open this email thread :-). Great to know you go with DB replication. It is our end goal as well. it is a intermediate state with BIND master -> PDNS -> BIND slaves currently. speaking of slaves, does PowerDNS Recursor can fully replace BIND in your experience? i.e. forwarding, caching and DNSBL, etc.? Thanks, > regards > Klaus > > _______________________________________________ > Pdns-users mailing list > [email protected] > https://mailman.powerdns.com/mailman/listinfo/pdns-users _______________________________________________ Pdns-users mailing list [email protected] https://mailman.powerdns.com/mailman/listinfo/pdns-users
