> As for technical solutions: > > * Archives that have not received a new message over a certain period > of time could be targeted for deletion. I am sure there are archives > that are no longer used, so they could be removed. The key is to > determine what is the proper period.
I'm already de-indexing them from the list of lists after six months of inactivity. Deletion of defunct lists is probably quite reasonable. > * Related to the previous one is to delete archives that have not > been accessed over a certain period of time. Let usage determine what > should and should not stay. Robot hits should be excluded. Some > heuristics may need to be employed since some robots do not play > nice (like address harvesters). Very hard. * Robot identification is hard * Robot traffic is high * Apache logs are enormously large (so I don't keep a long history) * I've turned off the "atime" records in the filesystem for improved performance. > * Remove archives that are just duplicates of a lists "official" > archives (this would actually affect me :-) For example, > I see that that are several cygwin.com lists archived at > mail-archive.com, but cygwin.com keeps there own set of archives > at <http://cygwin.com/lists.html>. Hard to identify. But banning all of YahooGroups was one step in this direction. > What is the space limitation of your current hosting provider? Current hosting provider is donating co-location service, but I can't swap out to a biggger machine. I have about 1.5 terabits there at the moment. -Jeff _______________________________________________ Gossip mailing list [EMAIL PROTECTED] http://jab.org/cgi-bin/mailman/listinfo/gossip