This comes from the conventional wisdom that backups of large databases must require high I/O throughput, something that LPARs and other virtual machines cannot provide. Basically all of that CW is wrong. Certainly, a 20TB data warehouse that needs to be able to be backed up in a few hours requires high levels of IO, but not every database has those short windows, nor are LPARs unable to support that amount of throughput. A heavily loaded physical system with a pile of LPARs _could_ have problems with throughput, but its not a truism that LPAR automatically equals lower throughput than a physical system, nor that a large data set means automatically high throughput requirements. Also, I seem to remember from the original post that this database is around 20GB - that hardly qualifies as large these days. Turn this into a numbers game - how much data, how often does it need to be backed up, and how long can it take to back up? Based on that, you can make an informed decision about whether an LPAR can support the IO. Matt ________________________________ From: oracle-l-bounce@xxxxxxxxxxxxx [mailto:oracle-l-bounce@xxxxxxxxxxxxx] On Behalf Of Allen, Brandon Sent: Tuesday, June 19, 2007 12:33 PM To: mkline1@xxxxxxxxxxx; oracle-l@xxxxxxxxxxxxx Subject: RE: Netbackup on a RAC We run 3 Oracle 10g environments on a single p570 split into 3 LPARs and they all backup to Netbackup with no problems. They aren't RAC, but that seems irrelevant to me. I'd ask your Netbackup guy for some substantial evidence to backup his claim before making any design decisions based on it.