> I've actually seen a scenario where intermittent RAM problems caused mostly > hidden differences in the two disk instances of a software raid mirror. It > wasn't fun to diagnose and fix since the disk reads would randomly have good > and bad data even after the RAM was fixed. I'm sure that is a rare thing, > but...
If you're using software/hardware mirror then you're already doing it wrong... ;) Unless RIAK has changed the rules they recommended raid 0 (because by definition RIAK is already redundant). Regardless of the hardware technology used, I fully believe that there will never be a 100% foolproof solution to a random cosmic event. Maybe when we're using quantum computers the rules might be a little different. I suspect that the basho crew can better answer how the hash is validate to determine if a node (or it's subsequent elements). _______________________________________________ riak-users mailing list riak-users@lists.basho.com http://lists.basho.com/mailman/listinfo/riak-users_lists.basho.com
