DNS Failover? - TCP-IP

This is a discussion on DNS Failover? - TCP-IP ; How do operations that demand extremely high reliability handle the issue of DNS failover? We use bind on a FreeBSD platform and it is incredibly robust but even the best systems fail or the communications gear fails as did the ...

+ Reply to Thread
Results 1 to 2 of 2

Thread: DNS Failover?

  1. DNS Failover?


    How do operations that demand extremely high reliability
    handle the issue of DNS failover? We use bind on a FreeBSD platform
    and it is incredibly robust but even the best systems fail or the
    communications gear fails as did the Ethernet switch port on one of
    our slave boxes this very day.

    How do they detect a true dns failure?

    The worst thing I can imagine happening is to have the
    mechanism for detecting a DNS failure trigger falsely so that it
    brings up the backup system on the same address as the working master.
    If there wasn't a DNS failure before, there sure will be one now.

    Thanks for any ideas.
    --

    Martin McCormick WB5AGZ Stillwater, OK
    Information Technology Division Network Operations Group

  2. Re: DNS Failover?

    In article ,
    Martin McCormick wrote:
    > How do operations that demand extremely high reliability
    >handle the issue of DNS failover? We use bind on a FreeBSD platform
    >and it is incredibly robust but even the best systems fail or the
    >communications gear fails as did the Ethernet switch port on one of
    >our slave boxes this very day.
    >
    > How do they detect a true dns failure?
    >
    > The worst thing I can imagine happening is to have the
    >mechanism for detecting a DNS failure trigger falsely so that it
    >brings up the backup system on the same address as the working master.
    >If there wasn't a DNS failure before, there sure will be one now.


    Authoritative DNS failover is handled automatically by listing multiple NS
    records for each zone.

    So I guess you must be talking about caching DNS. Resolvers allow you to
    list multiple nameserver, so they'll fail over automatically as well.
    However, this makes every DNS lookup slower, so it's generally considered
    just a stop-gap while you fix the problem with the first server.

    As an ISP, the way we deal with this is by having multiple, live servers
    responding to the same IP address. The routing protocol on our backbone
    automatically forwards the packets to the closest server. If a server goes
    down we remove the route, and the packets then go to another server.

    --
    Barry Margolin, barry.margolin@level3.com
    Level(3), Woburn, MA
    *** DON'T SEND TECHNICAL QUESTIONS DIRECTLY TO ME, post them to newsgroups.
    Please DON'T copy followups to me -- I'll assume it wasn't posted to the group.

+ Reply to Thread