Replacing servers in a cluster

In Platform Cluster Manager Community Edition, a system administrator can replace a physical server with a different or new physical server in the system. This is done from the Web Portal by updating the MAC address to reflect the newly replaced server. Platform Cluster Manager Community Edition ensures that the new physical server has the same server name and IP address as the old server. For example, if an existing server needs a hardware replacement, the replacement server can keep the existing configurations and network settings.

The impact of replacing a server that is in a cluster, is that any changes that were made to the server by a post-provision script are lost when the node is rebooted. This includes any configuration change or package installations. To get these changes back and have the server returned to its original state, the system administrator must rerun all post-provision scripts on the server.

Note: The system administrator only needs to rerun post-provision scripts on servers that use a stateful image profile. For servers that use a stateless image profile, the post-provision scripts are automatically rerun when the server is rebooted.

Before you begin

Make sure to completely shut down the old machine and unplug the server's BMC ethernet cable from the switch before you replace it with the new server.

Procedure

From the Web Portal, replace the server.
1. Log in to the Web Portal as a system administrator.
2. Select the Clusters tab, go to Clusters > Servers.
3. Select the server that you want to replace, and go to More > Replace.
4. Enter a new MAC address for the replacement server and click OK.
5. Click Close once the MAC address is updated.
Manually reboot the server to provision it. This can take a few minutes.
After the server is rebooted, rerun the post-provision scripts on the server. This step is only required if the server uses a stateful image profile, otherwise, skip this step.
1. From the command-line interface, go to the /opt/pcm/bin directory.
```
cd /opt/pcm/bin
```
2. Rerun the post-provision script on the server by issuing the pcm-run-cluster-scripts command. For example, if the server name is compute10 and the post-provison script is postComputeScript then issue the following command on one line:
```
# ./pcm-run-cluster-scripts –u root –p passw0rd -c Cluster01 -t ComputeTier –P postComputeScript –s compute10
```
  where Cluster01 and ComputeTier is the cluster and tier that server compute10 is found in.

Results

The cluster that the server is a part of is provisioned to include the replaced server. To view the state of the cluster, form the Web Portal, navigate to the Clusters tab, and select Clusters.