I've had some success turning off archiving as it reduces contention on the disk as well ensuring availability since I don't change back into one that's still archiving and hence not available. I did mention creating the indexes separately via script with nologging and parallel. On Wed, Oct 30, 2013 at 3:28 PM, Andy Klock <andy@xxxxxxxxxxxxxxx> wrote: > On Wed, Oct 30, 2013 at 4:12 PM, Storey, Robert (DCSO) > <RStorey@xxxxxxxxxxxxxxxxxx> wrote: > > No, its not portioned. > > > > I'm going to change the buffer size to reduce the commits and David > referenced doing the import with the db in noarchive mode. I could do the > import, then configure the setup and once I've verified that all is well, > place the db in archivelog mode before going live on it. > > > Running a database in "noarchivelog" mode doesn't reduce the amount of > REDO required for your import. Perhaps David meant to turn logging > off at the table level? However, imp classic doesn't do direct path > operations so you aren't going to get any savings there. Redo is for > recovery so personally I would leave it alone. > -- //www.freelists.org/webpage/oracle-l