This content has been marked as final. Show 5 replies
I'm not sure that I understand the question. Griddisks sit on top of celldisks. Also, all of your griddisks should be mounted in ASM already. Are you trying to reconfigure the Exadata storage? Maybe you should describe what you're trying to do.
Thank you. What I want to do is set physical independence among disks in different griddisks. My target is to isolate single point of failure (I know the matter of redundancy but I would prefer a true isolation among griddisks).
So are you looking to dedicate individual disks on a cell to specific disk groups?
If you're using ASM redundancy (which you would be crazy not to on Exadata), then you don't have a single point of failure, as the diskgroups are configured so that each cell is a failgroup. This means that 2 copies of an extent are not on the same cell. Keep in mind that to get the best performance on Exadata you need to spread your data across as many spindles as possible.
Thank you Andy, now it is clear. Do you recommend interleaving if it is planned several griddisks (4-5) or the default for data, recovery and dbfs?.
The default gives the faster outer tracks to data and the slower inner tracks to reco . This usually is the optimum configuration.