Dear All,

at HO I have a DNS servers (172.16.0.31)
and I have a DRC site..

I want to a DNS at DRC site .. all DNS records would be automatically synchronized to HO's DNS.

so If the HO DNS goes down then the DRC site
would take over .. could anybody please advise what should I do?

Please advise and Thanks

well you've two stages here, knowing that something has failed and then separately taking action. Depending on how this is all built, this is often two fundamentally unrelated problems to deal with.

HA DNS is (surprisingly?) hard, things like DNS TTL's make failovers frequently slow for one thing. Can you define what sort of level of environment this is in? Is this a spare box in your mates kitchen, or a multi million dollar data centre?

Thanks, our HO DNS is on the same network ..
HO and DRC site are connected with Leased-Line 100Mbps
HO DNS is on 172.16.0.31
and I want DRC DNS is on 172.16.0.131

Do you have an idea?

Thanks & Regards
Winanjaya

OK, still far far too little information to really base much on...

your servers on are the same /24 subnet?? Or is this two /25s? This is internal traffic, so presumably this is more than a single IP fail over? What tools do you have at your disposal? are you just looking at adding open source software to any conveniently placed servers, or do you have $50,000 to buy a couple of F5 GTM devices? What ARE these DNS servers?

Hi, yes they are on the same subnet /22
and they are Bind Named Server that come with Centos 5.5
and I am looking for Open Source softwsre that can help me to do DNS failover

I need advise

Thanks and Regards

well this article looks useful: http://www.freebsdwiki.net/index.php...over_A_records

this doesn't cover a active / standby scenario though, and a third site from which they are monitoring might not suit your model. You could easily rework the logic though I'm sure.

Hmm... but then... Are you really after using DNS to provide the failover, or are you looking at the failover of the DNS servers themselves? If it IS the latter, which feels increasingly likely, then why do you need to do it at all? what's wrong with a standard master / slave scenario? What do you think will happen from a failover? If they are synchonrized, then just use them both all the time... no?

hmmm. . master and slave?? I never try it... can the slave able to do takeover when the master goes down?
do I need to also modify the slave when the master was modified or when new record added?

please advise

Thanks and Regards

Winanjaya

there is no "takeover", they both just work all the time, the slave being told of any updates by the master. You can also look at a dual master set up if you really want.