IBM Tivoli Storage Manager, Version 7.1

Replicating data by command

You can replicate data that belongs to one or more client nodes or the data that belongs to a defined group of client nodes. To replicate data automatically and at a scheduled frequency, define an administrative schedule.

Procedure

You can replicate data in one of the following ways:

By issuing the REPLICATE NODE command on the command line
By using the DEFINE SCHEDULE command to create an administrative schedule, and specify when to start replication processing

To use an administrative schedule, you create it by using the DEFINE SCHEDULE command. When you start replication by using an administrative schedule, node replication processing occurs at a specified frequency. Regardless of how you start replication, the command must be issued on the source replication server.

For example, suppose that you have three client nodes, NODE1, NODE2, and NODE3. To replicate data from all nodes daily at 6:00 a.m. in a schedule named REPL_SCHED, issue the following command on the source replication server:

define schedule repl_sched type=administrative cmd="replicate node
node*" active=yes" 
description="replicate clients with name noden" starttime=06:00:00 period=1
perunits=days

Results

High-priority data is replicated before data with a normal priority. To schedule replication processes, you use the DEFINE SCHEDULE command.

If the data that belongs to a client node is being replicated, any attempt to replicate the data by issuing another REPLICATE NODE command fails. For example, suppose the backup data that belongs to a client node is scheduled for replication at 6:00 a.m. Replication of the archive data is scheduled for 8:00 a.m. Replication of the backup-data must complete before replication of the archive data starts.

Example

If you have many client nodes and are replicating a large amount of data, you can replicate data more efficiently by issuing several REPLICATE NODE commands in separate schedules. For example, replicate the data that belongs to the most important client nodes first in a single command. After the data that belongs to those client nodes is replicated, replicate the data that belongs to the other nodes.

Tip: To ensure that replication for first group of client nodes finishes before the replication for the other nodes starts, specify WAIT=YES on the first REPLICATE NODE command. For example, if you want to replicate the data that belongs to NODE1 and NODE2 before the data that belongs to NODE3 and NODE4, issue the following commands:

replicate node node1,node2 wait=yes
replicate node node3,node4

You can use the REPLBATCHSIZE and REPLSIZETHRESH server options to increase the speed of a node replication process between two replicating servers. The REPLBATCHSIZE option specifies the number of client files that are to be replicated in a batch, within the same server transaction. The REPLSIZETHRESH option specifies, in megabytes, a threshold for the amount of data that is replicated as a batch, within the same server transaction.

Feedback