Well, this is one case where I think that Oracle is being pretty good about telling you what's happening. I'd start by making sure that your spfile is shared between all instances. That will force the control_files parameter to be the same on all nodes (since it should be *.control_files).
Once you do that, if the problem still exists, then it is probably the case that your devices aren't mapped the same on this new node you're adding. That is /dev/rdsk/c2t0d59s6 isn't the same device from the other RAC nodes. You'll probably need a sysadmin to help determine which devices are which on each node. If you can (without harming data), I usually use dd to make sure that devices are the same on all nodes. That is:
if=/etc/hosts of=/dev/rdsk/c2t0d0s6 bs=4k
then, go to node2 and do:
node2$ dd if=/dev/rdsk/c2t0d0s6 of=/tmp/hosts-from-node1 bs=4k
and verify that /tmp/hosts-from-node1 matches /etc/hosts on node1. Of course, when you're adding a node to an existing cluster, you probably can't do that, but I thought it may be helpful. However, you could do this:
node1$ dd if=/dev/rdsk/c2t0d59s6 of=/tmp/fromnode1.dbf bs=8k count=100
node2$ dd if=/dev/rdsk/c2t0d59s6 of=/tmp/fromnode2.dbf bs=8k count=100
Then, ftp (in binary mode) or scp the files to the same node and do a diff on them. If they're different, then the devices aren't the same one on each node. There should be no harm to the data on this device by doing the dd outlined above, but I'd still do it when the existing RAC instances are shutdown (I wouldn't use abort in this case :).
From: Ujang Jaenudin <email@example.com>
Sent: Wednesday, July 4, 2007 9:17:52 AM
Subject: missing controlfile when add node
does anyone have an experience regarding add nodes in RAC environment
and about controlfile wrong?
3rd node was added, and try to mount database:
ORA-00203: using the wrong control files
ORA-00202: control file: '/dev/rdsk/c2t0d59s6'