Storage agent overview
Storage agent allows client systems to write data directly to, or read data directly from, storage devices that are attached to a storage area network (SAN). This is called LAN-free data movement.
LAN-free data movement makes LAN bandwidth available for other uses and decreases the load on the IBM Storage Protect server, allowing it to support a greater number of concurrent client connections.
A client system without the storage agent can send its data over the LAN (using TCP/IP) to a client with the storage agent. The storage agent client can then send the data LAN-free over the SAN to storage resources which the storage agent client shares with the IBM Storage Protect server. You can install the storage agent on a client system, which does not share storage resources with the IBM Storage Protect server. In this scenario, the storage agent will not be able to send data LAN-free; the data would have to go to the server using TCP/IP.
As shown in Figure 2, the storage agent can support several clients while installed on only one of the clients. You can also install the storage agent on a client system that does not share storage resources with the IBM Storage Protect server, but that is connected to a client system that does share storage resources. The LANFREECOMMMETHOD option set to TCP/IP allows a client system to send data to the storage agent. The LANFREECOMMMETHOD option also allows the storage agent to support several clients while the storage agent is installed on only one of the clients.
A IBM Storage Protect server, acting as a library manager, controls the storage devices. This server can be the server that is working with the storage agent or another IBM Storage Protect server in the enterprise. The IBM Storage Protect server tracks the metadata that the client has stored. The metadata, such as policy information, and file name and size, is passed over the LAN connection between the storage agent and server.
The storage agent communicates with the server to obtain and store database information, and to coordinate device and volume access. The server and client coordinate and transfer data access through the SAN. The client uses the storage agent for operations where appropriate. For example, if a SAN path is defined, the client (through the storage agent) transfers data using that path. If a failure occurs on the SAN path, failover occurs and the client uses its LAN connection to the IBM Storage Protect server and moves the client data over the LAN.
The storage agent can send the data directly to the server using the paths between the storage agent and the server. An example is a LAN-free storage pool that is updated to read-only after the client connects to the server and obtains its initial policy information. The storage agent, instead of failing the operation, sends the data to the server. If the storage hierarchy is configured so that the next storage pool destination is available, the server runs the operation.
You can also prevent data from being transferred over the LAN by specifying the IBM Storage Protect server parameters DATAREADPATH and DATAWRITEPATH with the REGISTER NODE or UPDATE NODE commands for a specific node. To review these settings, issue the following command on the server for the node:
query node node_name format=detailed
IBM Storage Protect supports SAN-attached device sharing in the following environments:
- IBM Storage Protect native library management support consisting of an ACSLS, SCSI, or IBM 349X library manager and library clients or just a library manager.
- Shared disk storage using a FILE library and the integration of IBM General Parallel File System, or IBM TotalStorage SAN File System. General Parallel File System is the preferred option for operating systems on which it is supported.
- External libraries.