FC SVM on AFF900 cluster 4 nodes

[ale]

Hi to all,

we have a cluster with 4 nodes on the same rack,. 

 AFF900-PA::cluster> cluster show
Node Health Eligibility
--------------------- ------- ------------
AFF900-PA-01 true true
AFF900-PA-02 true true
AFF900-PA-03 true true
AFF900-PA-04 true true
4 entries were displayed.

From gui we view cluster separated in 2 couple of group (view image):

G1 -> PA-01/PA-03

G2 -> PA-02/PA-04

We have defined an svm (only fc protocol) anc we create lif only on port 10a and 10b of physical port of G1.

I ask to you is it possible in case of disaster on G1 migrate automatically SVM_FC on G2?

If create SVM_FC with lif 10a and 11a of all 4 nodes, it is possible, in case of disaster of G1 migrate automatically SVM_FC on G2.

Thanks to all  

[TMACMD · Accepted solution]

When using SAN, it is a Best Practice to place ONE FC LIF from Each Fabric on EACH node.

2 node cluster = 2 LIFs per node = 4 FC LIFS

4 node cluster = 2 LIFs per node = 8 FC LIFS

You then do your single initiator zoning for each host to the NetApp.

Then if you allow the defaults to happen, Selective LUN Mapping (SLM) will advertise the node hosting the LUN and the HA-Partner FC LIFs to clients. If you need to move the LUN, ONTAP will update the advertisement to use the FC LIFs on the destination nodes.

Bottom Line, please place at least TWO FC LIFs (one per Fabric) on each node.