 |  |
Re: weird dns error resolving "www"
From: Len Conrad
Date: Monday, September 24, 2001
Time: 7:10:04 am
>that sound's really lame...the whole point of having multiple NS is so that
>if one goes down, queries for the domain will still resolve. What you're
>telling me is if an NS goes down, other NS's won't query the working NS and
>will just timeout.
the DNS "should" wait for the timeout, and then query the next NS. But
it's not only the DNS timing out, it's also the apps originating the
queries like browsers.
>..so what's the point of the redundancy?
"there's a lot of broke software out there"
>Again, my lack of knowledge about DNS is showing...is is something like if
>one NS gets an authoritative answer from another NS, it will continue to try
>that NS AND ONLY that NS until domain's TTL expires?
Such behaviour depends on the DNS software. eg, BIND8,9 will use RTT
calculations to home in, repeatedly, on the NS with the lowest RTT, round
trip time.
TTL has no effect on which NS is selected.
domain's don't have TTL, only resource records, and each can have its own
specific value.
once a record is in a DNS's cache, that DNS won't query the authoritative
DNS until after the record's TTL expires.
>I'm really trying to understand what's going on here. Thanks for your help
>to this point.
I've suggested a hypothesis (rednecks call this a "guesstimate"), not a
definitive analysis. I don't have any evidence other than what you say,
and how my dig commands responded.
I and Chris suggest you fix your authority pb, at least. Maybe in doing
that, you will have a happy secondary effect.
Having one NS permanently off-line will cause a "diminished user
experience" since 1/3 of www visitors whose DNS hits ns2 fist will have to
wait the timeout.
Len
Messages In This Thread:
Return to Digital Point Solutions' Home Page