Not as simple as it might first seem. A restore would involve multiple systems and likely many TB of data. On a system this large and with this kind of money going through, I would expect that they have data snapshots mirrored to another system that they could move over to pretty quickly rather than restoring the infrastructure. But even if they could roll over to redundant systems to say 10 minutes before the outage occurred, what about all the transactions in the DB that occurred after that point? Like the GEM purchase I completed an instant before the system crashed? Would create and nightmare for uCool to try to untangle. That is why they would rather suffer the loss of revenue from hours of down time than from minutes of data loss.Originally Posted by lionelmacroni_3982108
Restore would only happen if there was a catastrophic and otherwise unrecoverable failure of the system or DB.
