HP StorageWorks Storage Mirroring for Linux User's Guide (T2558-96078, February 2008)

6 - 3

Throughput Diagnostic Utility

Built into Storage Mirroring is a null, or non-existent, target that allows you to imitate a normal connection without passing

any data across the network. The purpose of this connection is to generate statistics that can be used in approximating the

time and amount of bandwidth that this connection will use when actively established. The name of the target is TDU or

Throughput Diagnostics Utility. The statistics are logged to a file. By default, the file is called

statistic.sts.For more

information on the statistics logging file, see DTStat on page 13-23.

Connection IDs

A unique connection ID is associated with each Storage Mirroring connection. The connection ID provides a reference point

for each connection and is used as an option in many of the DTCL commands.

Each connection is assigned an ID which is a sequential number starting at one. Each time a connection is established, the ID

counter is incremented. It is reset back to one each time the Storage Mirroring daemon is restarted. For example, if the

Storage Mirroring daemon was started and the same replication set was connected to five target machines, each connection

would have a unique connection ID from 1 to 5.

The connection IDs are displayed in the Management Console and the Text Client, but are not displayed in the Command

Line Client. If you are using the Command Line Client, you will need to keep track of the connection ID manually or establish

a variable so that the connection ID can be used with the DTCL commands. For detailed steps on creating variables, see

Variables on page A-39.

How Connections Are Established

To better understand how Storage Mirroring connections work, the following diagram identifies the sequential steps

completed when establishing a Storage Mirroring connection. Each numbered step is described below the diagram.

1. When the client initiates a connection, the Storage Mirroring source machine validates the data contained in the

replication set and the target information entered in the Connection Manager. The source machine verifies that if the

same source and target are connected through a different replication set, the specified target path is not already in use. If

the target path is in use by another Storage Mirroring connection, the potential for data overwrites exists. Therefore, if

one-to-one configuration is being used, Storage Mirroring will not allow a connection to the same path.

2. After the connection information is validated, the replication set database is sent to the target machine as a backup. The

file on the target machine is identical to the source’s machine name so that the file can be recovered if the source

machine crashes.

3. The source establishes transmission queues. Mirror and replication data is put onto these queues. For example, when

using a slow network or target machine, the data is updated on the source faster than it can be transmitted to the target.

The data is stored on a queue so that it is not lost.

4. A logical Storage Mirroring connection is established. This is the connection that is visible through the Storage Mirroring

client applications.

5. If mirroring or replication begins on connection, a network connection is established and the transmission of data begins

immediately. If no activity is set to occur on connection, the network link is established when mirroring and replication

are triggered.

Source

Target

Replication Set

Transmission

Queues

Replication Set

Backup