Originally posted by: arappoport
we are setting up a Tuning Manager (HTnM) infrastructure across a couple of Brocade Fabrics, and I would like to confirm the best design rules with the community.
We've 4 objects to manipulate:
HTnM Agent for SAN switch
Brocade SMI-A Host
Brocade switch fabric proxy
Each fabric stretches across 2 sites, HTnM is active at one site, deactivated at the other.
SMI-A allows you to configure multiple switch proxies per fabrics, which closes off the SPoF topic for the switch proxies.
But what about the HTnM Agent for SAN Switch and instances configurations?
HDS HTnM Installation guide recommends one SMI-A host per fabric, due to performance concerns if all fabrics performance metrics are being captured by one SMI-A host (p4-22)
And what is the best design for a simple DR plan? ie should the active HTnM Agent looks at the local SMI-A and in a DR situation, we reconfigure the HTnM agent and instances to point to the new SMI-A?
Many thanks in advance for your answers on this.