BIND and Cisco Local Director: problem with VIPs

Barry Margolin barry.margolin at level3.com
Wed Nov 12 18:52:52 UTC 2003


In article <botuo8$16o9$1 at sf1.isc.org>, Sara <demone33 at yahoo.it> wrote:
>This is my situation (IPs and hostnames dummy for privacy reasons):
>
>Using BIND 9.1.2 on Linux RedHat 7.3:
>Master name server:  pr_dns.mydomain.com  20.1.1.10
>Slaves:              sl_dns1.mydomain.com 20.1.1.11
>                     sl_dns2.mydomain.com 20.1.1.12
>
>I have Cisco Local Director configured in this way:
>
>VIPs: dns1.mydomain.com 20.1.1.101  ---> balanced to 20.1.1.11 and
>20.1.1.12
>      dns2.mydomain.com 20.1.1.102  ---> balanced to 20.1.1.11 and
>20.1.1.12

Why are you doing this?  Simply listing the two real addresses in the NS
records will normally cause them to be load-shared automatically.

>Problem: when I add a host to mydomain.com and increase the serial #,
>BIND tries to contact the NS record hosts! Therefore it may happens
>that the Local Director forwards all the packets to the same real
>hosts, thus preventing one slave from being updated!!!
>How can I solve this issue?

Use the "also-notify" and "notify explicit" options to send the Notify
messages to the real addresses instead of the virtual IPs.

-- 
Barry Margolin, barry.margolin at 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.


More information about the bind-users mailing list