Re: Reducing the 2560 negative cache default
Matthew Dempsky <[email protected]> Mon, 17 May 2010 14:15:38 -0700
| Newsgroups | gmane.network.djbdns |
|---|---|
| Message-ID | <[email protected]> |
On Mon, May 17, 2010 at 1:04 PM, Chris Pugh <[email protected]> wrote: > On 17 May 2010 19:49, Matthew Dempsky <[email protected]> wrote: >> The fields in the Z line go directly into the SOA record. (Also, >> dnscache never touches a data or data.cdb file directly.) > > Not quite what, > > http://cr.yp.to/djbdns/dnscache.html > > says? dnscache gets its DNS data from authoritative servers such as tinydns. tinydns gets its DNS data from data.cdb. data.cdb is generated from data by tinydns-data. Hence, dnscache doesn't touch data or data.cdb directly, so it's silly to ask how dnscache interprets lines in data. > dnscache does not cache SOA records. However, it does use SOA TTLs to determine > cache times (up to an hour) for zero-record responses and nonexistent domains. Yes, dnscache uses only the SOA TTL for negative caching, but other servers (e.g., ones that follow the negative caching RFC's advice) take into account the SOA record's "minimum" field.