Re: [Add] [EXTERNAL] Re: TTL of resolver.arpa

[Steffen Nurpmeso <[e-mail]>]

Tommy Jensen wrote in
 <[e-mail]\
 ook.com>:
 |I’m in full agreement with Ben on all points.
 |
 |Re-reading the draft, I see that the understanding that DDR designations \
 |are only valid for the network they were discovered on was left unspoken. \
 |I’ll open a new issue and prepare text to provide that guidance.
 |
 |As for the TTL topic, using a TTL is necessary for network configuration \
 |changes. If any given resolver doesn’t want to be polled, they are \
 |free to set a very large TTL. If a resolver wants to frequently confirm \
 |designations, they can set a very short TTL. Taking away that flexibility \
 |doesn’t make sense to me. As far as the meaning of TTL: it is obviously \
 |treated as a term of validity in the wild and not assuming that is \
 |the case is not practical.

Yes.  Here are configurable conf_(min|max|neg)_ttl variables in
order to allow overwriting of what DNS gives.  The former defaults
to 60 (1<60<120), (min<U32_MAX-1) the last to 60 (30<60<300,
talking about RFCs 1034, 1536 and 2308), pretty straightforward.
Special treatment for resolver.arpa provided addresses has to
happen (no file-based monitoring applies for them), but the lesser
the impact the better i would cite.

Thanks and ciao,

--steffen
|
|Der Kragenbaer,                The moon bear,
|der holt sich munter           he cheerfully and one by one
|einen nach dem anderen runter  wa.ks himself off
|(By Robert Gernhardt)