Managing Shared Disks

Converting twin-tailed virtual shared disks to concurrent virtual shared disks

Follow these general steps to convert twin-tailed virtual shared disks to concurrent virtual shared disks. The data on the virtual shared disks will remain intact.

1. Bring down any application that uses the virtual shared disks for the duration of this migration.
2. Stop the RVSD subsystem on all VSD nodes in the configuration. For example:

   ha.vsd stop
   

3. Update the vsdnodes to be part of a concurrent volume group cluster. For example:

   updatevsdnode -n primary_node,secondary_node -c cluster_name
   

4. Make the volume group concurrent capable. For example, on the primary node, issue:

   chvg -c volume_group_name
   

5. Export and import the volume group on both the primary and secondary nodes. For example:

   exportvg volume_group_name
   importvg -y volume_group_name hdisk
   

6. Disable auto varyon for the volume group on both the primary and secondary nodes. For example:

   chvg -a n volume_group_name
   

7. Delete the old volume group information. For example:

   SDRDeleteObjects VSD_Global_Volume_Group \
   global_group_name==your_global_group_name
   

8. Define the concurrent global volume group information. For example:

   vsdvg -g global_volume_group_name \
         -l primary_node:secondary_node volume_group_name
   

9. If the primary and secondary nodes were not previously part of a concurrent volume group cluster, they must be rebooted now.
10. Restart the RVSD subsystem on all VSD nodes in the configuration (the same nodes you stopped). For example:

    ha.vsd start
    

See the sample script /usr/lpp/csd/samples/vsdtocvsd.perl (which comes without warranty). This script will perform the previous steps for a varied off volume group.