IPMultiNIC[AB]? in parallel vs failover groups

[andyts]

We have recently started using IPMultiNICB with IPMP on Solaris 10. While reading documentation, I noticed that VCS guide example uses IPMultiNICB in a separate, parallel service group. Examples of IPMultiNIC and IPMultiNICA are both in a failover group. Anyone can explain why? In fact, I configured IPMultiNICB in a failover group (two-node cluster), and it appears that the resource is ONLINE on both nodes. Moreover, a different service group that references this resource via Proxy, can failover freely no matter where the group with IPMultiNICB happens to be. This makes sense, because Proxy, at least by default, has no notion of the location of TargetResName. So what is the difference? Or should I localize TargetSysName? Thanks, Andy

[Eric_Gao]

I don't see any example in putting IPMultiNICB in a parallel group,  but MultiNICB does.

Basiccly as MultiNICB is only used to test and get the state of interfaces, it is configured as parallel across the cluster.

[andyts]

Oops... My error.  Mental typo, I suppose :(  Of course, I meant MultiNIC[AB]? everywhere.  Sorry about that.  

But the basic question remains.  Why MultiNICB example is in parallel group, while MultiNIC and MiltiNICA are both in failover?  Is there a specific reason for the difference?

Andy