Scenario.
DNS config files are stored in a CVS repository and pushed out to
nameservers running BIND 8. Many of the domains have an NS entry that
needs to be removed. A typical zone file might look like
(let's say this is bogus.com.xx) << EOF
@ IN SOA server1.ns.zz. root.server1.ns.zz. (
2000090900
10800
3600
608400
86400 )
IN NS server1.ns.zz.
IN NS server2.ns.zz.
IN NS dead.ns.pp.
IN MX 50 mail
IN MX 100 another.server.
IN A 192.168.0.3
www IN A 192.168.0.4
mail IN A 192.168.0.8
EOF
The idea is to strip the ns entry dead.ns.pp and update the serial to
yyyymmdd30 (this being the serial format, last two digits are version for
that day, 30 can be considered to be safely high).
With this method one approach would be to cheak out each of the affected
zone files from CVS and update, then put it back in. The complication is
that not all zone files are in this format and there is no way to say that
any given file will be either known formats A, B or C.
I thought about doing silly things like axfr for the domains but I don't
even recall if the answers returned will be in a consistent format or not
(i could hunt around for a non-standard zone file to check against).
Looking for a bit of inspiration, anyone done something similar?
--
Alexander Else
http://cyberchrist.org
--
SLUG - Sydney Linux User Group Mailing List - http://slug.org.au/
More Info: http://slug.org.au/lists/listinfo/slug
