Generating a New Cluster

Below are the general procedures for generating a new cluster. For more information on using the Chef commands such as knife or Chef-client, see the Chef documentation.

1. Upload a cookbook to a server.

Upload a cookbook from a Chef Workstation to a Chef Server by using the knife cookbook upload.

2. Register run list.

Register a recipe to run list by using the knife node run_list add.

3. Install chef-client

Install the Chef client on the nodes in the new HA cluster using Chef. Perform the “knife bootstrap node name” command on the Chef Workstation to install.

4. Execute the recipe

Perform the chef-client command using the recipe registered at step 2 above via the run list command.

Execute the recipe on the node which will serve as the primary LifeKeeper node first. When that completes execute the recipe on the LifeKeeper standby node.

Notes:

• DataKeeper resource configurations – In some DataKeeper environments the device used must be added to the /opt/LifeKeeper/subsys/scsi/resources/DEVNAME/device_pattern file before the DataKeeper resource can be created. For more information on this refer to Known Issues and Restrictions or Troubleshooting.

• IP resource configurations – By default the IP resource health monitoring process attempts to determine if the virtual IP address can still send and receive packets outside the local node. This is done through broadcast pings to the subnet or via a ping of a specific IP address. If the device used by the template node is not reachable via the new cluster node, then this information must be updated. For more information refer to Viewing and Editing IP Configuration Properties.

5. Start the LifeKeeper GUI and confirm if each resource was generated as expected. At this point, all resource hierarchies are in the OSU state except DataKeeper.

6. Perform an in Service on the source node to start all protected resources.