The database is on the EMC storage with BCVs for both primary and standby. I can sync the BCV with the standby and mount the storage on a different host. Once I mount the storage, I can create one more physical standby database. Using the combination of this new physical standby, restore points, and flashback database, I can activate the standby and do some of our standard application checks to make sure the standby is valid. Once I am satisfied with the test, I can either put the database in standby mode or unmount the storage and establish the BCV back to the original standby. This entire process would take about 2-3 hours. I usually conduct this test often when time permits. In addition to this, we also have other tests that involve applications. In this way, I can mimic the real production environment and say that my DR plan is sound. Thanks, Senthil. On 10/2/07, Niall Litchfield <niall.litchfield@xxxxxxxxx> wrote: > > On 10/2/07, Senthil Subramanian <skumar.sen@xxxxxxxxx> wrote: > > > > Hi Carel-Jan, > > > > Thanks for your valuable input on this. > > I did try switching roles on a small 9i test database back in 2002, but > > never really implemented in a real production environment. > > We have a sound DR solution but never got a chance to test and implement > > the role transition. > > > > Please don't take this personally - since the above is a fairly common > thing to happen, but if you never really tested in a real environment then > how (unless you have recently survived a disaster) do you know you have a > sound DR solution. > > We have a 'sound' fire procedure where I work, when we actually tested it > about a year back we found it took 8 minutes to evacuate the (3 floor) > building and 20 to account for everyone. In other words testing the real > procedure in the real environment showed us some problems. DR is much the > same. You have to, regularly, test it for it to be worth the monetary > investment in setting it up. > > > > -- > Niall Litchfield > Oracle DBA > http://www.orawin.info