Howdy,

I'm running a simple Active/Passive IPMP setup with MultiNICB in IPMP
mode on a VCS 4.1 cluster. The MultiNICB resource is in a paralell
service group and I have IPMultiNICB's pointing to Proxy=>MultiNICB in
my groups.

I have IPMP configured with a test address on each NIC and a floating
administrative address (nodename) configured to start at boot. This
admin address is not managed by VCS.

After starting my service group, the IPMultiNICB address came up
on the Active interface but I noticed that my admin address was moved to
the Standby interface with the "Standby" flag set. I also noticed that
the IPMP group was changed to reflect the MultiNICB resource name. I had
to manually move the admin address over to the Active interface to get
back into the system.

# /etc/hosts
xxx.xxx.91.152alpha-ce2
xxx.xxx.91.153alpha-ce5
xxx.xxx.91.69alpha
xxx.xxx.91.63oracle

# /etc/hostname.ce2
alpha-ce2 netmask + broadcast + group service deprecated -failover up \
addif alpha netmask + broadcast + failover up

# /etc/hostname.ce5
alpha-ce5 netmask + broadcast + group service deprecated -failover
standby up

# ifconfig -a
ce2:
flags=9040843<UP,BRADCAST,RUNNING,MULTICAST,DEPRECATED,IPv4,NFAI LVER
mtu 1500 index 3
inet xxx.xxx.91.152 netmask ffffff00 broadcast xxx.xxx.91.255
groupname mnicb_multinic_91
ce2:1: flags=1000843<UP,BRADCAST,RUNNING,MULTICAST,IPv4mtu 1500 index 3
inet xxx.xxx.91.63 netmask ffffff00 broadcast xxx.xxx.91.255
ce5:
flags=29040843<UP,BRADCAST,RUNNING,MULTICAST,DEPRECATED,IPv4,NFAI LVER,STANDBY
mtu 1500 index 5
inet xxx.xxx.91.153 netmask ffffff00 broadcast xxx.xxx.91.255
groupname mnicb_multinic_91
ce5:1: flags=21000843<UP,BRADCAST,RUNNING,MULTICAST,IPv4,STANDBYmtu
1500 index 5 ^^^^^^^
inet xxx.xxx.91.69 netmask ffffff00 broadcast xxx.xxx.91.255

Anyone else seen this before?