I think that increasing the number of gossip peers only improves things. The probability of collisions I think does not systematically go up when you increase the number of peers, and it reduces the chance that you've chosen an unusually busy set of peers.
y On Sat, Aug 2, 2014 at 4:06 PM, Pete Stephenson <p...@heypete.com> wrote: > Hi all, > > I was reviewing the mailing list archives recently and had a few questions: > > 1. From what I've read, generating DB stats block the DB and SKS does > not respond to search/submit queries while the generation is ongoing. > > What happens to queries during the blocking period? Are they rejected or > dropped? Do they hold waiting for the DB to come back online? > > In my case, the server is using Apache as a reverse proxy. It uses an > SSD for storage, so stats generation takes <5 seconds or so. Is it > problematic to generate stats more frequently than once per day? > > 2. SKS can only gossip with one peer at a time. Additional gossip > traffic made while the server is gossiping with another peer fail and > must be retried. > > Is there some optimal number of peers that a server should have that > would balance between increasing the degree of interconnection and > gossip collisions? Is there a critical number of peers beyond which > collisions become an issue? > > Cheers! > -Pete > > > _______________________________________________ > Sks-devel mailing list > Sks-devel@nongnu.org > https://lists.nongnu.org/mailman/listinfo/sks-devel > _______________________________________________ Sks-devel mailing list Sks-devel@nongnu.org https://lists.nongnu.org/mailman/listinfo/sks-devel