When configuring CF by CLI, perform the following steps.
In this section, the cluster system configured with two nodes where the CF node names are "fuji2" and "fuji3", is explained as an example.
Create CIP configuration files.
Specify /etc/cip.cf as below on all the nodes which configure the cluster system.
Example:
fuji2 fuji2RMS:netmask:255.255.255.0 fuji3 fuji3RMS:netmask:255.255.255.0
Note
If you manually create /etc/cip.cf, you cannot reconfigure CF by Cluster Admin. To reconfigure CF by Cluster Admin, delete the /etc/cip.cf file beforehand.
Set IP addresses.
Specify /etc/hosts as below on all the nodes which configure the cluster system.
Example:
<cip address1> fuji2RMS <cip address2> fuji3RMS
Enable remote access by using cfcp/cfsh.
Specify /etc/default/cluster.config as below on all the nodes which configure the cluster system.
CFCP "cfcp" CFSH "cfsh"
Edit /etc/default/cluster on all the nodes.
Edit /etc/default/cluster to create the files of the following contents:
nodename <CF node name> clustername <Cluster name> device <Cluster interconnect 1> device <Cluster interconnect 2>
Example:
nodename fuji2 clustername PRIMECLUSTER1 device eth2 device eth3
Note
Make sure that the node name to be defined in nodename is the CF node name, not the node name of the OS.
Set the owner, group, and access permission.
# chown root:root /etc/default/cluster # chmod 600 /etc/default/cluster
Reboot the nodes.
Execute the following command with any node in the cluster system and set the Cluster Integrity Monitor (CIM).
# rcqconfig -a <nodename> ...
nodename : CF node name
Example:
# rcqconfig -a fuji2 fuji3
If this command fails, check again that CF node names and cluster names configured in /etc/default/cluster in step 4 are correct.
Check that it can be communicated with the RMS node name.
Example: When checking from fuji2
# ping fuji3RMS
If it cannot be communicated, check again that CF node names, RMS node names, and CIP addresses configured in /etc/cip.cf and /etc/hosts in step 1 and step 2 are correct.