I converted a single instance to RAC manually. I re-created the control
file, I added & enabled an additional logfile thread, additional undo
space and everything works as needed:

Connected to:
Oracle Database 11g Enterprise Edition Release 11.2.0.1.0 - Production
With the Partitioning, Real Application Clusters, OLAP, Data Mining
and Real Application Testing options

SQL> select * from v$active_instances;

INST_NUMBER
-----------
INST_NAME
--------------------------------------------------------------------------------
1
lpo-oracle-01:vmsp1

2
lpo-oracle-02:vmsp2



There is only one little problem:
-bash-3.00$ srvctl status instance -i vmsp1 -d vms_p
PRCD-1027 : Failed to retrieve database vms_p
PRCD-1035 : vms_p is the database unique name of a single instance
database, not a cluster database
-bash-3.00$

How can I explain to clusterware what did I do? I am reluctant to open a
TAR, it will take days. Does anybody know off hand?


--
http://mgogala.byethost5.com

On Feb 24, 5:41*pm, Mladen Gogala <n... (AT) email (DOT) here.invalid> wrote:
Hi Mladen,

I have just started testing on 11gr2 (so this may not work), but have
you tried...

srvctl add instance -d db_name -i instance_name -n node_name

If that doesn't work, can you afford the downtime to drop the database
definition in the clusterware and re-add it with

srvctl add database -d db_name -o $ORACLE_HOME

HTH,

Steve

On Thu, 24 Feb 2011 16:05:56 -0800, Steve Howard wrote:



Thanks, Steve! I did my reading and opened a TAR never the less. Yes,
this was the solution. I had to remove a single instance DB first, but
that was not a problem.


--
http://mgogala.byethost5.com