Re: Negative Caching
Patrick Domack <[email protected]>
| Newsgroups | gmane.network.dns.powerdns.devel |
|---|---|
| Message-ID | <[email protected]> |
Quoting Peter van Dijk <[email protected]>: > Hello Patrick, > > On Nov 2, 2012, at 17:10 , Patrick Domack wrote: > >> I have been adjusting the source to increase the negative cache >> size, for when quering against rbl's >> >> From my stats, it seems to me that the negative cache is not part >> of the positive cache, but totally different cache area/counters. >> >> So it seems the real max-cache-size = max-cache-size * 110% (as the >> negative cache is hard coded at 10% of the max-cache-size currently). >> >> I adjusted my source, so I could specify the negative cache size as >> a percentage of the max-cache-size, but as it seems to be >> different, I'm wondering if my patch (before I submit it in a >> feature request) should be changed to a negative-cache-size, that >> would just default to the current behavure. > > > Concrete numbers are usually better than percentages of other numbers. > > In any case, please submit your patch at wiki.powerdns.com so people > can look at it! > I'll submit it sometime next week, I've been playing with it, and seeing how it works, it does seem to be seperate from stats point of view, and I had decided next time I look at it, to change it from percent to just a raw number, and just leave the default as a percent if undefined, so there isn't any suprises on an upgrade.