I think you are saying redo instead on undo. 800GB/day is quite a lot. It depends on whether your i/o subsystem can sustain this throughput. This can be found out from the storage vendor. Secondly the redo write rate depends at the rate at which the application commits. Batch commits are much better. Another option is trying use use unrecoverable loads, with selective backup of tablespaces maybe by using split mirror. RAID 10 is recommended. For this redo rate, it is better to allocate separate set of disks only for redo. Regards Joseph Amalraj Edgar Chupit <chupit@xxxxxxxxx> wrote: On 1/6/06, Deepak Sharma wrote: > Is RAID 10 ok for placing Redo Logs? > > It's for a very high write-intensive database, > generating 800G of Undo/Day. > > > > -- > //www.freelists.org/webpage/oracle-l > > > -- Best regards, Edgar Chupit callto://edgar.chupit -- //www.freelists.org/webpage/oracle-l