Cluster Manager GUI cannot connect after specifying IP address

[khemmerl]

Hello everyone.  I was going through the configuration of a service group for an Oracle database using the Cluster Manager GUI.  I had duccessfully defined the disk groups and volumes and had just got to the IP address.  I have two servers: sunblade1 (xx.xx.xx.211).and sunblade2 (xx.xx.xx.212).  I created a resource "sunblade_IP" of type "IP" and set the following values:

Device:
> Per System:
> > sunblade01: e1000g0
> > sunblade02: e1000g0
Address:
> Per System:
> > sunblade01: 123.123.123.211
> > sunblade02: 123.123.123.212

Unfortunately, as soon as I clicked OK my Cluster Manager GUI died and it is not able to reconnect.  I've tried to get the Cluster Manager GUI to connect to the servers (sunblade01 and sunblade02), the IP address of the server (xx.xx.xx.211 and xx.xx.xx.212), the cluster name (suncluster01) and the IP address of the cluster (xx.xx.xx.213) but nothing works.  My questions are:

1. What information is supposed to go under the IP Resource?  I was told it is supposed to be the IP addresses of you servers.  Should it be the IP of the cluster?
2. Now that I've lost this use of the GUI, how do I fix this?

Thanks in advance for any help.

Ken

[mikebounds · Accepted solution]

Ken,

IP resource in VCS is for virtual IPs - i.e IPs that you want to be able to move between the 2 servers.  As you entered physialc IPs you probably messed up your networking.  You should delete your resource:

 hares -delete sunblade_IP 

Sort out your networking using manual ifconfig commands and then you should be able to get back into GUI.

You don't HAVE to have a cluster IP - this is only if you want to manage (i.e log into the JavaGUI) using virtual cluster IP.  But you will need a virtual IP for Oracle service group that your Oracle clients use.

Mike

[khemmerl]

As luck would have it, the change I made to the IP address killed the connection between the GUI and the servers before I could do a File > Save.  The result is that the IP address configuration never made it into the main.cf file and a simple reboot of both node of my cluster has resolved the problem without the need for me to use the command you supplied.

Thanks for the quick response Mike.

Ken