The objective is to move the DB from the SAN onto "local" drives, sorry, should have made myself clear. Dan Norris wrote: > Why clone the DB? Since you're not using ASM, I presume you're using > OCFS (hopefully not RAW). > > So, converting is just shutting down the 2nd instance, setting > cluster_database=FALSE, and dropping the unnecessary redo log (thread > #2) and undo tablespace. That wouldn't require any changes to the > dataguard configuration since the database name remains the same and > the downtime should be minimal. Some updates to tnsnames.ora for the > clients and you should be all set. Maybe I missed something in your > request, but I don't see why this would require a clone. > > Dan > > On Sat, Jan 31, 2009 at 1:09 PM, Tony Sequeira <tony@xxxxxxxxxxxxxxx > <mailto:tony@xxxxxxxxxxxxxxx>> wrote: > > Oracle 9.2.0.6.0 EE - I know > Win 2003 Server SE SP1 > > I've been asked to look at converting the the 2 node RAC, to a single > instance database, running on node 1 of the old RAC system. SAN > hardware getting old, replacement and spares difficult to find. There > must be a fall back to the original RAC DB if performance suffers. > > My initial thoughts are to clone the database onto a new instance, and > change the global tnsnames.ora, so the client application will not > notice the difference. > > However, there is a dataguard standby, which will also be impacted, > which will need to some kind of action, whatever is done (I believe). > >