On Fri, Sep 23, 2011 at 3:49 AM, Tommy.Fallsen@kongsberg.com wrote:
Would like input from the list how others have done it.
Hi Tommy,
Will you be using Ontap flexclone, snapshots or replication? Those are all at the volume level, so arch, redo, temp and data files should be on separate netapp volumes. With NFS, you have an upper limit on ESX 4.1 of 64 NFS mount points, so it might be used up fast if you have a lot of instance, especially if you clone/snap/replicate...
Our vsphere-hosted Oracle uses dNFS in 11g to access NFS volumes directly using the guests virtual network. The ESX hosts are all redundant 10G. We found VMXNET3 to be a very efficient virtual network adapter that was within error margin of the physical machine we compared it do.