Logical host resource in SC3.2 missing IPMP

Question

I was under the impression that when creating a logical host resource in Solaris Cluster 3.2, that it will read the hosts and hostname.* files to get the IPMP info and do this automatically for you. I am not seeing this behavior and am hoping someone can shed some light on this for me. Here is the relevant data, any help is greatly appreciated. If it matters, this is Solaris Cluster 3.2 on Solaris 10, running in a 25k domain.
# grep cluster1vip hosts
10.230.108.28 cluster1vip cluster1vip.example.com
10.230.108.29 cluster1vip-ce1 #ipmp_test_cluster1vip1
10.230.108.30 cluster1vip-ce5 #ipmp_test_cluster1vip2
# cat hostname.ce1
cluster1vip netmask + broadcast + group ipmp_sc1 up \
addif cluster1vip-ce1 deprecated -failover netmask + up
# cat hostname.ce5
cluster1vip-ce5 deprecated -failover netmask + broadcast + group ipmp_sc1 up
When I bring the resource online, these are the settings that get applied:
ce6:1: flags=1040843<UP,BROADCAST,RUNNING,MULTICAST,DEPRECATED,IPv4> mtu 1500 index 3
inet 10.230.108.28 netmask ffffff80 broadcast 10.230.108.127
Again, any help or guidance is greatly appreciated.

Answer 1

For probe based IPMP the following has always worked for me:
- IPMP interfaces of same media type
- SPARC needs local-mac-address set to true in OBP
- Supported interface (if using probe method - [http://docs.sun.com/app/docs/doc/816-4554/mpoverview?a=view|http://docs.sun.com/app/docs/doc/816-4554/mpoverview?a=view] )
/etc/hosts
192.168.1.1 hawkeye
192.168.1.2 ipmp-test1
192.168.1.3 ipmp-test2
For the cluster node 'hawkeye'
/etc/hostname.ce0 before the IPMP change:
hawkeye group sc_ipmp0 -failover
/etc/hostname.ce0 after the IPMP change:
hawkeye group sc_ipmp- failover up addif ipmp-test1 -failover deprecated up
/etc/hostname.ce.1
group sc_ipmp0 failover up addif ipmp-test2 -failover deprecated up
Reboot cluster node and you should be good to go.