Hi there,
in the Sanity.pm's there are checks for the entered ttl's - they should be >= 300 and <= 2,592,000.
For Nameserver-Records and for the Zone itself I'm fully d'accord. But why should single RRs not have a ttl smaller than 300? RFC's (1034, 1035, 1794) say that they can be 0 - which would cause them to be non-cached.
I have a customer asking for this feature.
Which reason is there they are limited to be >= 300 (besides BIND's known irritation)?
BTW, Matt: My customers using NicTool here are very pleased with it. Good work!
ciao,
Dorothea
The reason for the lower limit is client interaction. If you publish records with TTLs lower than 300, there's quite a number of DNS resolvers that ignore the low TTL and then set their own when they cache the record. That's sort of self-defeating and it severely deprecated the value of low TTLs.
Now, it may be that the older DNS implementations that misbehaved in such ways have all been upgraded and it's perfectly safe to do this now. I'd say, give it a try and see (if you haven't already).