3/10    What's "unix clustering"?
        \_ Short answer: having a bunch of machine with same config and some
        \_ Short answer: having a bunch of machines with same config and some
           'clustering software' that makes it so these machines are all
           redundant backups for each other and also share horsepower for
           things like cpu, io, etc.  There's more than one type of clustering
           and quality, price, features, etc vary widely between vendors,
           versions, etc, but this is the gist of it for laymen.
           \_ Mainstream products that do "clustering" are almost
              always failover systems.  Parallel processing is much
              rarer.  Sun and Veritas both have clustering products.
                -brianm
              \_ clustering >> failover.
                 a cluster normally handles "failover", but a failover
                 solution does not neccessarily do clustering.
                 A [group of] system[s] that is purely a "failover system"
                 is NOT a "cluster"
                 \_ Guys, I was trying to just get the main idea across and
                    not start a lengthy shitfest over the details.  Thanks.
        \_ For me clustering involves non-trivial degree of communication
        between the nodes, ie tightly-coupled cluster vs. a simple
        server farm. Most people use the two terms interchangeably, though.
                                -muchandr
        \_ Thanks, much appreciated!
        \_ something that any stupid academic fool understands.
                \_ Troll?  Cookie?  Mmmmmm!  Good troll!  Yummy!