Adding Compute Nodes to the Cluster
Applies To: Windows Compute Cluster Server 2003
After creating the head node, you will add one or more compute nodes to your cluster. Adding and removing nodes can be performed during initial setup from the To Do List or by using the Compute Cluster Administrator.
You can add a compute node to your cluster using the following methods:
Automated Addition — This method installs an operating system and all required Compute Cluster Pack software on the node, then adds the node to the cluster. To add a compute node using this method, refer to Using Automated Addition to Add Nodes.
This method utilizes Remote Installation Services (RIS) from the head node. To use the Automated Addition method, the cluster network topology must include an intra-cluster private network. If the cluster is installed only on a public (organizational) network, you cannot use Remote Installation Services. If you do have a private network, RIS must be installed and configured on the head node before you can use this method. Also, you must have already created and modified your RIS installation images with valid software license keys.
Compute Cluster Pack Installation method — This method can only be used on servers that have a supported operating system installed. After you have run Compute Cluster Pack Setup, the Compute Cluster Installation Wizard presents the option Join this server to an existing compute cluster as a compute node. At this time you can:
Specify a cluster head node. If you specify an accessible head node, the node is added to the cluster.
Choose not to specify a head node or, if you do, the specified head node is not found. If you do not specify a head node or a specified head node is not found, the new compute node is not joined to any cluster. You can join the node to a cluster at a later time by using the Manual Addition method. This is useful if you want to prepare compute nodes in advance before you have determined which cluster they will join.
To use this method, refer to Installing Compute Cluster Pack on a Compute Node
Manual Addition — This method can be used only after a supported operating system and the Compute Cluster Node Manager Service are installed and running on the prospective node. The computer must also be added to the same Active Directory domain as the head node. You must be a local administrator on the computer and it must be attached to a network. You must know the computer name of any node that you want to add to your cluster. To use this method, refer to Using Manual Addition to Add Nodes
If you deploy compute cluster nodes by using non-Microsoft disk imaging products, make sure that the node disk image has been prepared using the Microsoft System Preparation tool. Refer to How to use the System Preparation Tool (sysprep.exe) to Perform Disk Duplication (http://go.microsoft.com/fwlink/?LinkId=63514).
After a compute node is added to a cluster, the default status is Pending Approval. At this time, you can run scripts, install applications, set environmental variables, or perform other administrative tasks as needed.
In those cluster topologies where the compute nodes are isolated behind the head node (cluster topology scenarios 1 and 3), if you want the compute nodes to have access to the Internet through a proxy server on the public network, you must specify a proxy server on each compute node. This can be accomplished by the use of scripts, Group Policy, or manually at the node. Use the following procedure to specify a proxy server manually.
To specify a proxy server manually
Open Internet Explorer
Click Tools, then click Internet Options.
On the Internet Options dialog, click the Connections tab, then click LAN Settings.
Unselect Automatically detect settings and click OK.
On the Connections tab, click Settings, then specify the proxy server. You might need to specify a fully qualified machine name (for example, servername.domain.company.com)