This content has been marked as final. Show 2 replies
We have had many customers that use a second Exadata rack for mixed-use DR and dev/test. If they're using dataguard for DR (which most are), they don't worry too much about instance caging. Processing on the dataguard side is pretty minimal. In the event of a disaster, some of the dev/test databases would be shut down, depending on how hard they're hitting the system.
I always prefer to use RAC for both the primary and standby side. In the event of a disaster, you're going to want the standby to behave as similar to production as possible. We have done this with Exadata racks of all sizes, ranging from quarter racks to X2-8 systems.
We use our Exadata V1 (yes, V1) as a Development machine and we also have our Production Standby (Data Guard) instance on there for DR. Our Production machine is a V2.
There's no need for instance caging as there is minimal consumption for the Data Guard instance. However, when we switch over to run Production on the V1, it's on the assumption that we may have to shutdown the Development database. The difference in memory resource between V2 and V1 is significant, so there's no expectation that Development would stay running at that point.
N.B. there are no Flash cards on the V1, so the performance is noticeably reduced on certain batch jobs when we run Production on the V1