Round Robin Compatibility with RDM \ Claim rule settings

[kwg66]

OK, brand new cluster \ production environment just built, v6.5d,  just set the policy for the default SATP to use RR for all LUNs visible, rebooted each host, and as expected all multiple-pathing to the LUNs has been changed form fixed to RR.   (esxcli storage nmp satp set --default-psp VMW_PSP_RR –satp VMW_SATP_DEFAULT_AA)

Before migrations begin to this new environment, Need to verify that the MSCS and SQL clusters using RDMs aren't going to break with the RR configuration.

Can someone please point me to a compatibility reference regarding RDMs and RR?

If RR is not compatible with the RDMs in use, then I need to understand the end result of the claim rule put in place that sets all LUNs for the default SATP to RR. 

- if I manually change RR back to Fixed for the RDMs, will that persist after reboot given the rule put in place? 

[dskwared]

We have Round Robin multipathing configured as the default (including RDM LUNs) in our environment, and I don't recall seeing any issues as a result of that configuration.

[Finikiez]

If your storage array is compatible with RR then you can use it with disk attached as RDM into VM as well.

There is no difference how disk device is used as RDM or as VMFS datastore.

[kwg66 · Accepted solution]

Thanks for responding, but I needed a bit more, and I have now found the answer to my question.   As of vSphere 5.5, RR can be used with MSCS clusters that leverage RDMs. 

See here for the details:   VMware Knowledge Base