New to NetApp, recently receivedtwo-node cluster - AFF300. Implementation was done by consulting company, using existing FC infrastructure.
I have been asked to retire the FC infrastructure (Brocade FC switches), and move our VMWare workloads to NFS.
I have deployed NFS as proof of concept, using one interface, but now I would like to go production.
Each node has 2 10 Gb NiCs, each connected to a separate Nexus data swtich. No VLANs, or custom IP spaces are configured. Reading through documentation, implementation best practices, etc, I have come up with the following:
1. create Interface group in dynamic mode for each node
2. create one LIF per node
3. create broadcast domain for all NFS interfaces
4. configure Necus and VMWare as per best practices.
I am posting this under "Ask the Experts" to see if design is sound. Any comments, suggestions are highly appreciated.
- if your switches can do it, utlize vPC between the switches for the LACP.
- Like you mentioned about broadcast domains, isolate the NFS as much as posiable. own broadcast domain and even on it's own private non-routeable VLAN. If you can't do a segmented off vlan, just be sure to set your export polices correctly.
- I would create 1 lif per production VMware datastore. That way the lif can follow the volume around if you ever move the volume to a different node in the cluster.