9.2 Recovering the Original Cluster Configuration at Node Expansion

When the original cluster configuration must be restored due to a fault or interruption during node expansion, take the following steps for recovery:

Procedure

Cancel the configuration changes in GFS.
If a new node was added to the GFS configuration at node expansion, it is necessary to cancel the GFS configuration in multi-user mode before deleting the new node.
Execute sfcnode on any one of the cluster nodes and delete the node configuration information. Then, execute sfcfrmstop and stop GFS on all the nodes. In this state, execute sfcsetup on the node to be deleted, and then delete the node information from the management partition.
See
For information on how to use each command, see the "PRIMECLUSTER Global File Servers Configuration and Administration Guide."
Change the class scope.
If a new node was added to the GDS configuration, it is necessary to change the class scope by executing the sdxattr command. For information on the change procedure, see the "Global Disk Services Configuration and Administration Guide."
Recover the PRIMECLUSTER system configuration files in single user mode.
See "Chapter 11 Backing Up and Restoring a PRIMECLUSTER System."
Recover the GLS configuration files.
Restore the environment definition files that are used by the GLS redundant line control by executing the following command.
```
node1# /opt/FJSVhanet/usr/sbin/hanetrestore -f /var/tmp/backup/<save file name>
```
Use "hanetYYYYMMDD.bk" as the save file name.
For details on the hanetrestore command, see the "PRIMECLUSTER Global Link Services Configuration and Administration Guide: Redundant Line Control Function."
Restore the environment definition files used by the GLS multipath function by using the following procedure:
```
node1# cd /etc/opt/FJSVmpnet
node1# uncompress /var/tmp/backup/mpnetfile.tar.Z
node1# tar xvf /var/tmp/backup/mpnetfile.tar
```
Restore the original cluster configuration.
Restore the cluster interconnect, LAN cable, and RCI cable to the original cluster configuration. At this time, restore the original RCI settings. This should be done by field engineers.