HP Serviceguard for Linux Version A.11.19 Deployment Guide, September 2012

c. Deactivate the volume group:

umount /dev/vgws/lvol1

vgchange -a n vgws

vgchange –-deltag $(uname –n) vgws

2. Repeat the steps on the second server.

You should see the content from each server in the test file.

Configure the Cluster

In the next steps, you will create the cluster, define the node membership, configure the cluster

heartbeat and cluster lock LUN device.

1. From an internet browser such as Internet Explorer, invoke HP System Management

Homepage, https://[hostname]:2381. For example:

https://eve.cup.hp.com:2381

2. Login (use root user and password set during installation of the operating system).

3. Go to the Tools tab.

4. Click the “Serviceguard Manager” link to launch the Serviceguard Manager.

5. Click the “Create Cluster” button on the right.

NOTE: You may see the following message: “There is no cluster configured.”

6. In the Create Cluster window, enter the Cluster Name, for example “Test” and

enter checkmarks in the boxes for both nodes, for example: adam, eve.

7. Go to the Network tab,

Enter in the “Subnets” section, for example:

Subnet: 16.89.84.128, Type: Heartbeat

Enter in the “Select Subnet Configuration” section, for example:

Node Network Address

adam bond0 16.89.84.245

eve bond0 16.89.84.247

8. Go to the Lock tab

For the Cluster Lock Type, Select “Lock Lun”

Enter the Lock Lun Path for each node, for example:

Node Lock Lun Path

adam /dev/mapper/mpath0p1

eve /dev/mapper/mpath0p1

Select Finish.

NOTE: When using Device Mapper Multipath, the path to the cluster Lock LUN, for example

/dev/mapper/mpath0p1, must be the same on each node.

9. Select Check Configuration Look for any errors

NOTE: You may get a warning about the default NODE_TIMEOUT value. This warning can

be ignored here, but refer to the documentation when finalizing your cluster.

10. Select Apply Configuration. Select “OK” in the pop-up dialog box.

11. To verify the cluster configuration, run the following options from Administration menu of the

HP Serviceguard Manager Summary page to test that each node can run the cluster in the

event that the other node fails:

a. Administration -> Run Cluster (on both nodes)

b. Administration -> Halt Node (select adam)

c. Administration -> Run Node (on adam)