Hi
Thanks Very Much to these guys, James Sainsbury, Steve VanDevender, JEFFREY
THOMAS, Hudson James, Johan.Brusche for their help and hints.
My question:
I have 1 primary server & 2 secondary servers running and configured in
tru64 4.0F with the latest PK7.
Problem:
When the primary server was shutdown (or powered off) , from the DNS
client machine I typed, nslookup <hostname> or ping , it took a long
time to respond, more than a minute which is not good at all as I assume
the secondary server must take over to resolve request straight away.
But From the BIND secondary server, I got a response straight away when
I either ping or nslookup any hosts.
Below is their responses
Typical resolvers use the list of name server IP addresses in
/etc/resolv.conf in the order listed. If the first does not respond
after some timeout period, then the second is tried, then the third,
etc. The timeout is application-configurable but typically not in a way
you can easily reconfigure.
If your resolv.conf lists your primary and secondary nameservers in that
order, and you know the primary is going to be down, temporarily
removing the primary from resolv.conf will cause the secondary to be
queried immediately.
-----------------------------------------------------------
I would guess that I has to do with the DNS timeout on the client, that
is how long the client will wait for the response from the primary
before trying the secondary... You didn't state the OS on the client.
Seems like Windows waits a very long time before trying the secondary.
I've had applications timeout before the secondary was queried. Does
anyone else have any more info on this?
------------------------------------------------------------------
Bind is 'passive'. It merely listens for requests. If the primary bind
server is
down, the secondary does not 'take over'. The client machines must be
configured to
look at the next DNS server. On the clients add the secondary servers to
the /etc/resolv.conf file.
-------------------------------------------------------------
The retry count and interval coded into nslookup
are set such that the failover to the secondary
server happens after 75 secs.
You have some control over it by using the environment
variable RES_OPTIONS
RES_OPTIONS="retrans:3 retry:2 "
See manpage resolver or do a google on RES_OPTIONS
Alan.Nguyen_at_au.transport.bombardier.com
Received on Fri Jan 17 2003 - 02:13:43 NZDT