Ora Fail Safe - under cluster ctl on one node but gives FS:10413 on other

Question

Hi,
Trying to set up a single Fail Safe (3.3.3) instance (9.2) on a set of Windows 2003 servers running MSCS.
After creating the standalone DB on the first node, I was able to get it verified and added to the cluster control (while the second server was down - had to this as the error I'm about to mention kept rolling back the changes on both nodes, otherwise). Now that the first node is configured and the FS listener service and FS DB service are created, I tried setting the second node to match the first one (viz. getting listener.ora, sqlnet.ora and even tnsnames.ora).
The "verify group" operation kept failing on creation of the Windows FS service for the DB and listener.
Found out that there was an issue with picking up the Oracle_Home when even trying to invoke sqlplus manually from the command line (though the registry entries seemed to be right and match the ones set on the first node). Eventually on adding the ORACLE_HOME variable to the list of env variables, it started picking up the OH correctly and the next "verify group" operation successfully added the WIndows FS service for the DB (OracleService{SID}) on the second node. (The first node didn't have to encounter these problems as during the standalone DB install (using DBCA), this service was created).
Next, started the Windows FS DB service (on the second node with the problem) and tried starting the DB manually (out of cluster controll) and it worked fine.
However when trying to "verify group" again, I keep getting the error FS-10413 (and the subsequent FS:10046, FS:10111 and FS:10890) while trying to add the Windows FS listener service (OracleORA_HOMETNSListenerFsl{oradb_virtual_name}).
Could really do with extra eyes (and heads) helping to get to the bottom of this. Thanks.
<<< The gist of the output from the "verify group" action is below >>>
13 18:29:18 > FS-10300: Verifying Oracle Net listener resource OracleORA_HOME1TNSListenerFsloradbl
14 18:29:19 Oracle Net listener Fsloradbl for the group is not found
FS-10413: Error found in Oracle Net on node PANTHER; the group listener needs to be taken offline. Do you want it to be fixed? > Yes
15 18:29:19 >> FS-10600: Oracle Net configuration file updated: C:\ORACLE\PRODUCT\9.2\NETWORK\ADMIN\LISTENER.ORA
16 18:29:20 ** ERROR : FS-10066: Failed to start Windows service OracleORA_HOME1TNSListenerFsloradbl for the Oracle Net listener
<<<< >>>>>>>
Further to typing the above, read the section 7.10.8.2 in "Oracle Fail Safe Concepts and Admin guide" (http://download-uk.oracle.com/docs/html/B12070_01/ofs_db.htm#sthref985) and checked the output log. Bizarrely, the output file has a few "msg not found" errors (akin to when it doesn't find the correct OH).
Wonder where this has to be set in the registry!?!
The TNS-01151 (missing listener name in listener.ora) error is also a bit odd as this is the new listener OFS is trying to create and add to the new listener.ora file. Isn't it!?!?!
<<<< Fsloradbl.out >>>>>
LSNRCTL for 32-bit Windows: Version 9.2.0.7.0 - Production on 07-FEB-2007 21:50:46
Copyright (c) 1991, 2002, Oracle Corporation. All rights reserved.
Starting tnslsnr: please wait...
TNSLSNR for 32-bit Windows: Version 9.2.0.7.0 - Production
Message 279 not found; No message file for product=NETWORK, facility=NL
TNS-01151: Message 1151 not found; No message file for product=NETWORK, facility=TNS     [Fsloradbl]
Listener failed to start. See the error message(s) above...
<<<< >>>>
Thanks for your time (reading and hopefully responding to this) and any comments or suggestions.
Regards,
Ranjit

Answer 1

yes i did. i mean, the db starts corretly using the Enteprise Manager and listener.ora and tnsnames.ora are the ones created during db creation. they seem alright.
i'm trying to execute those commands from the node of the cluster where the db resides.
but... i have one big question: do i have to create the db instance on BOTH cluster nodes (both pointing to the same oradata in the shared disk) or is it enough to create my standalone db only on the nodeA and then let everything to the FailSafe management (adding the standalone db to the cluster group FailSafe creates all the services needed)???
thanks