On October 27, 2002 at 21:40, Jeff Breidenbach wrote: >* Server: > > Good news. We seem to be running quite smoothly on the new server > "poet", and it looks like all the network stuff has transitioned > over. Yet another drive failed two days ago, but the RAID-10 array > merely laughed and kept on running. The old machine "zamboni" will > be taken off line this week.
Did you get second-hand disk drives? It seems odd to have two new drives fail in a short-period of time. Who is the manufacturer? >* Money: > > The colocation folks mentioned above is a genuine 501(c)(3) > non-profit organization. That means people now have the option of > donating money to them to offet Mail-Archive's bandwidth costs, see > FAQ[1] for details. My guess is the service will continue to be > primarily Jeff-funded, but let's see what happens. How do they charge you? Is it by the bandwidth you use, or is it a flat cost? BTW, in your FAQ question, "Why are you using this engineering design?" You left another alternative approach: Each list to be archived is manually approved. I.e. You still have one archiving address account, archive@jab.org, but you would have to explicitly add each list to be archived instead of having scripts attempting to auto-detect new lists. The known lists would then be used in the filtering processing as new mail comes in (avoiding the need for complicated heuristics in the scripts). Now, this more administrative interaction approach could be improved by have a submission form for users to submit the list address they want archived. The form then auto-adds the list to the known-lists database (probably a text file) for use in the filtering. This avoids human admin interaction to add a new list, but provides an avenue where addition of lists must go through an approval process in case the policies of mail-archive.com ever change. --ewh _______________________________________________ Gossip mailing list [EMAIL PROTECTED] http://jab.org/cgi-bin/mailman/listinfo/gossip
