ICRV Sync

Each SP participates in cluster management in either a Primary or Secondary role based on a number of criteria. ICRV Sync is handled by proprietary mirroring functions within the CPC's.

Split Brain situation is prevented by utilizing various mechanisms
1) Having two independent hardware links between peers
2) Proprietary logic to handle situations where a double fault causes both peers to believe they are in control. This logic includes detection and resolution of split-brain based on past and current state information that each peer had received from the other node and the current state of each of the hardware links between the peers. These are just a couple of various data points that are considered in identification and resolution.

I'm surprised to read that the mirroring is performed by the CPC and not the DPC. Does this means that the ICRV is only updated when changes is done at the CPC ? I expected that configuration changes was updated on the primary ICRV and the DPC would mirror this to the secondary ICRV.

Can you please clarify what method is used to avoid a split brain situation ?
I'm a former HACMP expert and even here this was a challenge, and this is the same for MSCS and all other cluster solutions.
I know that the IBM SAN Volume Controller use a quorum device to control this and therefore you can't have a two site implementation with only one SAN Volume Controller installed. This will give a false security.

I think you are confusing normal virtualization mirroring function (storage elements that are virtualized) with ICRV mirroring which are handled differently. Virtualization mirroring is handled by the DPC, with reporting and handling of any control issues back to the CPC nodes. ICRV's are handled directly by the CPC not DPC. Split brain is handled as indicated in the previous response.