Configuring client/server functions

This section is relevant only if you installed CS Linux to run in a client/server environment (with multiple CS Linux nodes in the same network).

Many resources, such as ports and LUs, are configured on an individual node. These are known as node resources.

Other resources, are common to all nodes; only one definition for the resource is maintained for the entire domain. Such resources are known as domain resources. Domain resource definitions are stored only on the controller server for the domain, and are accessible from all the nodes in the domain.

Note: A standalone CS Linux system has only one server; that server always acts as the controller.

In a client/server environment, a server can be marked as a configuration server; CS Linux maintains a list of these configuration servers. The first server listed is the controller server, and any other servers listed are backup servers. The servers are listed in order, so that the second server listed (the first backup server) takes over if the controller server is unavailable, the third server listed (the second backup server) takes over if neither the controller nor the first backup server is available, and so on.

When any of the nodes in the domain are active, the first available configuration server in the domain (the first server that can be contacted and has CS Linux software running) becomes the controller server. If the current controller becomes unavailable (because it cannot be contacted, perhaps due to a network failure, or because the SNA software running on it is stopped), the next available configuration server in the list becomes the new controller.

CS Linux can run without a controller. This happens if none of the servers in the configuration server list can be contacted. If this happens, you can view and configure node resources only on the servers that can be contacted.

Note: You cannot directly indicate which node acts as the controller server; the controller server is selected based on the order in which nodes are added to the configuration server list. If you wish to move a server to the top of the list, remove all other nodes from the list and then add them again.

You can also use the following administration commands to query, add, and delete configuration servers:

query_sna_net: Lists the servers in the file.
add_backup: Adds a new server to the end of the list.
delete_backup: Removes a server from the list. You can use the delete_backup command to delete either the controller server (so that the second server listed takes over as controller) or a backup server (so that it can no longer act as the controller).

Note: You cannot delete a server if it is the only server listed on which the CS Linux software is running, because in this case there is no other server that can take over as the controller server. At least one enabled controller server is required in a client/server configuration.

CS Linux uses the local hostname setting when it starts up to identify the node. It may be required to use a DNS alias name instead. To set the CS Linux server's name to a DNS alias name, set the following statement:

export SNA_SERVER_NAME=DNS_alias_name_for_server

in the environment file /etc/opt/ibm/sna/environment.

Clients are available for Windows, AIX and the different Linux platforms. In addition, the client can run in a container or AIX WPAR.

Managing CS Linux client/server systems provides information about advanced Client/Server configuration, including how to move clients and servers into different CS Linux domains and how to configure the details of client operation.