Simple two node Cluster Install - Hung after reboot of first node

Question

Hello,
Over the past couple of days I have tried to install a simple two node cluster using two identical SunFire X4200s, firstly following the recipe in: http://www.sun.com/software/solaris/howtoguides/twonodecluster.jsp
and when that failed referring to http://docs.sun.com/app/docs/doc/819-0912 and http://docs.sun.com/app/docs/doc/819-2970.
I am trying to keep the install process as simple as possible, no switch, just back to back connections for the internal networking (node1 e1000g0 <--> node2 e1000g0, node1 e1000g1 <--> node2 e1000g1)
I ran the installer on both X4200s with default answers. This went through smoothly without problems.
I ran scinstall on node1, first time through, choosing "typical" as suggested in the how to guide. Everything goes OK (no errors) node2 reboots, but node1 just sits there waiting for node2, no errors, nothing....
I also tried rerunning scinstall choosing "Custom", and then selecting the no switch option. Same thing happened.
I must be doing something stupid, it's such a simple setup! Any ideas??
Here's the final screen from node1 (dcmds0) in both cases:
Cluster Creation
Log file - /var/cluster/logs/install/scinstall.log.940
Checking installation status ... done
The Sun Cluster software is installed on "dcmds0".
The Sun Cluster software is installed on "dcmds1".
Started sccheck on "dcmds0".
Started sccheck on "dcmds1".
sccheck completed with no errors or warnings for "dcmds0".
sccheck completed with no errors or warnings for "dcmds1".
Configuring "dcmds1" ... done
Rebooting "dcmds1" ...
Output from scconf on node2 (dcmds1):
bash-3.00# scconf -p
Cluster name: dcmdscluster
Cluster ID: 0x47538959
Cluster install mode: enabled
Cluster private net: 172.16.0.0
Cluster private netmask: 255.255.248.0
Cluster maximum nodes: 64
Cluster maximum private networks: 10
Cluster new node authentication: unix
Cluster authorized-node list: dcmds0 dcmds1
Cluster transport heart beat timeout: 10000
Cluster transport heart beat quantum: 1000
Round Robin Load Balancing UDP session timeout: 480
Cluster nodes: dcmds1
Cluster node name: dcmds1
Node ID: 1
Node enabled: yes
Node private hostname: clusternode1-priv
Node quorum vote count: 1
Node reservation key: 0x4753895900000001
Node zones: <NULL>
CPU shares for global zone: 1
Minimum CPU requested for global zone: 1
Node transport adapters: e1000g0 e1000g1
Node transport adapter: e1000g0
Adapter enabled: no
Adapter transport type: dlpi
Adapter property: device_name=e1000g
Adapter property: device_instance=0
Adapter property: lazy_free=1
Adapter property: dlpi_heartbeat_timeout=10000
Adapter property: dlpi_heartbeat_quantum=1000
Adapter property: nw_bandwidth=80
Adapter property: bandwidth=70
Adapter port names: <NULL>
Node transport adapter: e1000g1
Adapter enabled: no
Adapter transport type: dlpi
Adapter property: device_name=e1000g
Adapter property: device_instance=1
Adapter property: lazy_free=1
Adapter property: dlpi_heartbeat_timeout=10000
Adapter property: dlpi_heartbeat_quantum=1000
Adapter property: nw_bandwidth=80
Adapter property: bandwidth=70
Adapter port names: <NULL>
Cluster transport switches: <NULL>
Cluster transport cables
Endpoint Endpoint State
Quorum devices: <NULL>
Rob.

Answer 1

I have found out why the install hung - this needs to be added into the install guide(s) at once!! - It's VERY frustrating when an install guide is incomplete!
The solution is posted in the HA-Cluster OpenSolaris forums at:
http://opensolaris.org/os/community/ha-clusters/ohac/Documentation/SCXdocs/relnotes/#bugs
In particular, my problem was that I selected to make my Solaris install secure (A good idea, I thought!). Unfortunately, this stops Sun Cluster from working. To fix the problem you need to perform the following steps on each secured node:
Problem Summary: During Solaris installation, the setting of a restricted network profile disables external access to network services that Sun Cluster functionality uses, ie: The RPC communication service, which is required for cluster communication
Workaround: Restore external access to RPC communication.
Perform the following commands to restore external access to RPC communication.
# svccfg
svc:> select network/rpc/bind
svc:/network/rpc/bind> setprop config/local_only=false
svc:/network/rpc/bind> quit
# svcadm refresh network/rpc/bind:default
# svcprop network/rpc/bind:default | grep local_only
Once I applied these commands, the install process continued ... AT LAST!!!
Rob.