Re: Getting a consistent copy
Date: Thu, 6 Aug 2009 23:22:42 -0700 (PDT)
Few dozen times as that DB is their lifeline for the business. I was involved from the design. The db grows at around 20TB a day (also partitions are dropped around 15TB/day), and archivelogs will be around 60-80TB/day if the DB is in force logging mode. Most of the operations (loads) are done with no logging. So running the database in archivelog mode practically makes no sense.
Their backup is just to take a snapshop at SAN level by fencing the IO and mount that cloned/snapshots to another node and start the database to verify the validity of the backup. Technically it is treated like an instance failure and usual rollforward+rollback brings the database to a consistent state. The storage is managed by ASM and DB is multinode RAC if that matters.
-Gopal Received on Fri Aug 07 2009 - 01:22:42 CDT