Managing Shared Disks

Removing a node from a concurrent virtual shared disk configuration

To remove a node from a concurrent virtual shared disk configuration, do the following:

1. Shut down the applications on all the nodes that use virtual shared disks.
2. If the IBM Recoverable Virtual Shared Disk subsystem is running on any node, issue the following command:

   ha.vsd stop
   

3. On the control workstation, delete the global volume group information, then issue the following command:

   SDRDeleteObjects VSD_Global_volume_Group \ 
      global_group_name==your_global_volume_group_name
   

4. Define the global volume group to be a virtual shared disk rather than a concurrent virtual shared disk. To do this, issue the following command:

   vsdvg -g global_volume_group_name volume_group_name primary_node
   

5. On the primary node, update the volume group timestamps that the IBM Recoverable Virtual Shared Disk subsystem maintains. To do this, issue the following command:

   vsdvgts -a volume_group_name
   

6. Remove the number of concurrent volume groups defined from the node you are deleting. To do this, issue the following command:

   SDRChangeAttrValues VSD_Cluster_Info node_number==node_being_deleted \ 
      cvgs_defined=0   
   

7. On the control workstation, remove the nodes from the cluster. To do this, issue the following command:

   updatevsdnode -n node_number -c NONE
   

8. Start the RVSD subsystem on all virtual shared disk nodes. To do this, issue the following command on each node:

   ha.vsd start
   

9. Make all the nodes aware that a node has been deleted. To do this, issue the following command:

   ha.vsd refresh
   

10. Restart all applications.