Error 485680 reservation warning(node_join) - Unable to lookup local_only f

Question

SETUP
3xT5220
1xStorageTek 2540 FC RAID
SC 3.2
OS 10 Patchlevel 137111-08
Two of the 3 cluster nodes are directly connected to the RAID.
Multipathing enabled on all cluster nodes (stmsboot -e).
PROBLEM
On the node, which was added to an already established cluster as the third node, that is NOT connected to the RAID, I'm seeing hte following messages:
Cluster.CCR: [ID 485680 daemon.warning] reservation warning(node_join) - Unable to lookup local_only flag for device dsk/d10.
Cluster.CCR: [ID 485680 daemon.warning] reservation warning(node_join) - Unable to lookup local_only flag for device dsk/d9
svc.startd[8]: [ID 652011 daemon.warning] svc:/system/cluster/globaldevices:default: Method "/usr/cluster/lib/svc/method/globaldevices start" failed with exit status 96.
with exit status 96.
svc.startd[8]: [ID 748625 daemon.error] system/cluster/globaldevices:default misconfigured: transitioned to maintenance (see 'svcs -xv' for details)
Cluster.CCR: [ID 916112 daemon.error] /usr/cluster/bin/scgdevs: Filesystem /global/.devices/node@3 is not available in /etc/mnttab.
# svcs -xv
svc:/system/cluster/globaldevices:default (Suncluster globaldevices service)
State: maintenance since January 12, 2009  2:38:26 PM EST
Reason: Start method exited with $SMF_EXIT_ERR_CONFIG.
   See: http://sun.com/msg/SMF-8000-KS
   See: /var/svc/log/system-cluster-globaldevices:default.log
Impact: 3 dependent services are not running:
        svc:/system/cluster/mountgfs:default
        svc:/system/cluster/clusterdata:default
        svc:/system/cluster/ql_rgm:default
svc:/system/cluster/scsymon-srv:default (Sun Cluster SyMON Server Daemon)
State: maintenance since January 12, 2009  2:38:41 PM EST
Reason: Start method exited with $SMF_EXIT_ERR_CONFIG.
   See: http://sun.com/msg/SMF-8000-KS
   See: /var/svc/log/system-cluster-scsymon-srv:default.log
Impact: This service is not running.
# cat /var/svc/log/system-cluster-globaldevices:default.log
[ Jan 12 02:37:52 Disabled. ]
[ Jan 12 02:37:53 Rereading configuration. ]
[ Jan 12 02:47:55 Rereading configuration. ]
[ Jan 12 02:48:28 Enabled. ]
[ Jan 12 02:49:07 Executing start method ("/usr/cluster/lib/svc/method/globaldevices start") ]
[ Jan 12 02:49:10 Method "start" exited with status 96 ]
[ Jan 12 06:14:43 Executing start method ("/usr/cluster/lib/svc/method/globaldevices start") ]
[ Jan 12 06:14:46 Method "start" exited with status 96 ]
[ Jan 12 09:50:26 Executing start method ("/usr/cluster/lib/svc/method/globaldevices start") ]
[ Jan 12 09:50:34 Method "start" exited with status 96 ]
[ Jan 12 13:17:27 Executing start method ("/usr/cluster/lib/svc/method/globaldevices start") ]
[ Jan 12 13:17:30 Method "start" exited with status 96 ]
[ Jan 12 14:28:54 Executing start method ("/usr/cluster/lib/svc/method/globaldevices start") ]
[ Jan 12 14:29:02 Method "start" exited with status 96 ]
[ Jan 12 14:38:23 Executing start method ("/usr/cluster/lib/svc/method/globaldevices start") ]
[ Jan 12 14:38:26 Method "start" exited with status 96 ]This prevents my scalable resource group from coming online on the third cluster node.
From the cluster error message guide:
+485680 reservation warning(%s) - Unable to lookup local_only flag for device %s.+
Description: The device fencing program was unable to determine if the specified device is marked as local_only. This device will be treated as a non- local_only device and nodes not within the cluster will be fenced from it.
Solution: If the device in question is marked marked as local_only and is being used as the boot device or the mirror of a boot device for a node, then that node may be unable to access this device and hence, unable to boot. Contact your authorized Sun service provider to determine whether a workaround or patch is available.
Any idea what's going on or how to prevent it?

Answer 1

I have isolated the issue to ssh and the login window. I can use these same administrator credentials to initiate ARD sessions, configure the server via Server admin or WGM, authenticate to perform privileged finder operations.
restarts do not change anything.
per a related post, but not identical investigated /var/db/dslocal/nodes/Default/config/ I only had one of the referenced files, not the several hundred those posters seemed to have. I made a backup, and removed it, followed by a restart. There was no effect.
I opened Server Admin, and turned off ssh, verified that remote login was not checked in System Preferences. Logged out, no effect, restarted again no effect.
There is no joy with trying Local Administrator.
This is a clean 10.5.4 Install (Monday, August 3) DNS works per Anthony Rocco's setup instructions and verification methods. DNS was working prior to upgrading to OD Master. Kerberos logins worked properly. I was going through and setting up share points and setting ACLs when I realized I lost access.
I would take any suggestions people might have...
thanks,
Ion