ONTAP Discussions

ONTAP Select 9.15 - Cluster network must remain non-routable?

borkp
183 Views

I am deploying a 2-node ONTAP Select (9.15) cluster with the Deploy OVA.  The customer is seeking Metrocluster SDS functionality.  

 

As I understand, a 2-node cluster is a Metrocluster SDS deployment as long as the OTS nodes are on different ESXi hosts, and hosted on different VMware datastores.  Aggregates appear to be mirrored (plex0/1), so the complete loss of one full side will be survivable.

 

My question is on e0c/e0d for the Cluster network.  The default network is 169.254.x.x, which is a non-routable network.  My 2 basic questions are:

a. Is my understanding of a 2-node HA deployment correct?

b. Does this limit OTS to either a single datacenter, bridged network, or a stretched VLAN because of the private Cluster network?

Thank you.

 

1 ACCEPTED SOLUTION

chamfer
98 Views

 

@borkp  I am not sure whey NetApp changed the image for ONTAP Select MetroCluster SDS on the documentation here https://docs.netapp.com/us-en/ontap-select/concept_usecase_mc_sds.html

 

The Image below or here"https://www.netapp.com/media/main-DR3_tcm19-39080.png" show L2 connectivity between sites.

chamfer_1-1733895705663.png

 

To answer your questions:

 

a. Yes your understanding of a 2-node HA deployment is correct, each site needs locally survivable hardware (hosts+datastores).  Don't forget the mediator at the third site (site#3).

b. Yes, you are correct L2 typologies are the limitation, it also limits the distance between sites due to the 5ms RTT + 5ms Jitter requirements (5ms RTT with a maximum jitter of up to 5ms, for a total of 10ms maximum.latency).

 

 

 

View solution in original post

1 REPLY 1

chamfer
99 Views

 

@borkp  I am not sure whey NetApp changed the image for ONTAP Select MetroCluster SDS on the documentation here https://docs.netapp.com/us-en/ontap-select/concept_usecase_mc_sds.html

 

The Image below or here"https://www.netapp.com/media/main-DR3_tcm19-39080.png" show L2 connectivity between sites.

chamfer_1-1733895705663.png

 

To answer your questions:

 

a. Yes your understanding of a 2-node HA deployment is correct, each site needs locally survivable hardware (hosts+datastores).  Don't forget the mediator at the third site (site#3).

b. Yes, you are correct L2 typologies are the limitation, it also limits the distance between sites due to the 5ms RTT + 5ms Jitter requirements (5ms RTT with a maximum jitter of up to 5ms, for a total of 10ms maximum.latency).

 

 

 

Public