Re: Who admins DNS?
От | Magnus Hagander |
---|---|
Тема | Re: Who admins DNS? |
Дата | |
Msg-id | 6BCB9D8A16AC4241919521715F4D8BCEA0FB97@algol.sollentuna.se обсуждение исходный текст |
Ответ на | Re: Who admins DNS? ("Dave Page" <dpage@vale-housing.co.uk>) |
Ответы |
Re: Who admins DNS?
Re: Who admins DNS? |
Список | pgsql-www |
> > So, just to toss out another option, I'm pretty sure we could get a > > free account from the folks at UltraDNS for the postgresql.org > > project. This would give us something that is global, > replicated, w/ > > no-downtime, anycast capable, on a completely independent > > network/hardware system and accessible by whichever > community members > > we deem necessary. > > If you're thinking of them providing secondary services, then > there isn't really any need - it was the blind primary that > got screwed up in the snafu this week, the secondaries (and > in fact the visible primaries) were working just fine. For > added redundancy though, I'm configuring up a couple more > secondaries later this week. Agreed - as a secondary, the help is marginal. > If you're thinking of using them as a primary, then that is > highly unlikely to work I suspect. Our main zone file is > dynamically generated from the database on borg and a couple > of other sources on a daily basis > - it's not something that we can just edit in a web interface > periodically. There's also the mirrors.postgresql.org > subdomain which is handled by a different primary and is > dynamically updated every 15 minutes or so. Even if we left > that as-is, do UltraDNS provide a method to delegate the > subdomain to a different server? I'm *sure* they provide delegation of subdomain, given the customers that they have. AFAIK, UltraDNS is a professional *enterprise* DNS hoster. But I can be wrong on that ;) The more interesting question to ask is do they have an API for dynamically updating the DNS records. (Maybe just dynamic-dns? If security is configure dproperly, that should work, no?) If it's an API we can use, then it can definitly still be a win. //Magnus
В списке pgsql-www по дате отправления: