A friend asked: "Could you share what is the best practise for appling rolling patches
on database, grid infrastructure and Storage cells on EXADATA environment, we
are using half rack node in our environment?"
My reply:
You can have a look at Enterprise Manager's Provisioning and Patch Automation pack,
which can automate the application of rolling patchs on database and Grid Infrastructure.
It is worthwhile to spend a few extra dollars on Enterprise Manager packs to aid
with managment of complicated Exadata machines - and EM managment of Exadata is
going to get better and better! Get the customer to agree on the management
software cost (EM). It will definitely help them in the long run. EM will be
able to centrally manage Exadata machines in the organization, find performance
issues, patch and upgrade the database, etc.
In reality, CRS/RAC patching is used extensively by Enterprise Manager Grid Control customers for RAC databases. We have had good success stories in the real world. In Exadata, there are mostly 8 nodes in the RAC database so It makes sense to use the automated patch deployment procedures in the Provisioning and Patch automation pack of Grid Control when we can have rolling patch upgrades automatically handled by the deployment procedure. It is not rocket science.
Remember to apply the OMS cumulative patches first, of course. Core DBAs should remember that this is a way to ease their manual work and their late nights, especially if CPU/PSU patches are to be applied every three months (as per Oracle recommendations) to the RAC databases. Fundamentally, give it a whirl. If the pre-reqs pass, the patching will work. If not, you can either fix the issues or do the patching manually if you prefer.
