Re: rman parallel processes

I haven't used configured backups but I will try and answer it from the
point of view of how I have managed backup performance for large
databases. I use three key indicators

1. Agreed/Advertised SLAs on MTTR
2. Load placed during the read that constrict other concurrent activity
3. Available resources during the write

I tend to lean towards keeping the write pipeline reasonably full without
introducing thrashing and stay within those norms with a bit of head room
for unforseen growth.  For example you could use mkfile to test the
capabilities of the write target

-Krish

--
http://www.freelists.org/webpage/oracle-l

Other related posts: