Because HiRDB Datareplicator definitions and other values are updated when a unit or back-end server configuration is modified, you must stop HiRDB Datareplicator before performing such a change. The following shows the procedure for modifying the configuration of a unit or back-end server.

Procedure

1. Use the pdrplstop command to disconnect HiRDB Datareplicator.
2. Use the HiRDB Datareplicator hdestop command to stop the data extraction facility.
3. If Y is specified in the pd_rpl_init_start operand, change it to N.
4. Change the configuration of the unit or back-end server, and execute the system reconfiguration command. For details about changing the configuration of a unit or back-end server, see 11. Modifying the System Configuration.
5. Reconfigure the extraction environment of HiRDB Datareplicator.
6. Use the HiRDB Datareplicator hdestart command to start the data extraction facility.
7. Use the pdrplstart command to connect HiRDB Datareplicator.
8. If you changed pd_rpl_init_start to N in step 3, change N back to Y, and then execute the system reconfiguration command.

(a) Standby system switchover facility

Execute the system reconfiguration command only when the primary system is the same as the running system. In addition, use the system reconfiguration command only to change HiRDB system definitions on the primary (running) system. Change HiRDB system definitions on the secondary system by copying the HiRDB system definition files from the primary system to the secondary system. The following procedures show how to modify HiRDB system definitions when the standby system switchover facility is being used.

Procedure 1: When your cluster software is HA Monitor and the standby system switchover facility is in the server mode:

1. Execute the system reconfiguration command on the primary system. This causes the secondary system HiRDB to stop automatically.
2. After the system reconfiguration command processing finishes, copy the HiRDB system definition files from the primary system to the secondary system.
3. On the secondary system, execute the pdstart command (pdstart -q for a HiRDB parallel server configuration) to start the secondary system HiRDB.

Procedure 2: When Hitachi HA Toolkit Extension is being used:

1. Stop the secondary system HiRDB.
2. Execute the system reconfiguration command on the primary system.
3. After the system reconfiguration command processing finishes, copy the HiRDB system definition files from the primary system to the secondary system.
4. On the secondary system, execute the pdstart command (pdstart -q for a HiRDB parallel server configuration) to start the secondary system HiRDB.

Procedure 3: When the standby switchover facility is in the monitor mode:

1. Execute the system reconfiguration command on the primary system.
2. After the system reconfiguration command processing finishes, copy the HiRDB system definition files from the primary system to the secondary system.

(b) Standby-less system switchover facility

The following procedures show how to modify HiRDB system definitions when the standby-less system switchover facility is being used. Note that the system reconfiguration command cannot be used when the system has alternated (while the alternate BES is the active system).

Procedure 1: When the cluster software is HA Monitor:

1. Copy the following HiRDB system definitions into the $PDDIR/conf/chgconf directory in the alternate BES unit:
   Back-end server definitions for the normal BES
   Unit control information definitions for the normal BES unit
   
2. Execute the system reconfiguration command.

Procedure 2: When Hitachi HA Toolkit Extension is being used:

1. Copy the following HiRDB system definitions into the $PDDIR/conf/chgconf directory in the alternate BES unit:
   Back-end server definitions for the normal BES
   Unit control information definitions for the normal BES unit
   
2. Use the pdstop -q -c command to release the wait status of the alternate portion.
3. Execute the system reconfiguration command. Note that executing the system reconfiguration command automatically places the alternate portion whose wait status was released in step 2 back into wait status.

(c) Notes (for both standby and standby-less system switchover facilities)

Note 1 applies when the standby system switchover facility is being used, and note 2 applies when the standby-less switchover facility is being used. Note 3 applies for both facilities.

1. Because the HiRDB system definitions on the primary system do not match the definitions on the secondary system while the system reconfiguration command is executing, the system cannot be switched over by means of a planned system switchover or as a result of a server failure.
2. Because the back-end server definitions and the unit control information definitions on the normal BES unit do not match the equivalent definitions on the alternate BES unit while the system reconfiguration command is executing, the system cannot be switched by means of a planned system switchover or as a result of a server failure.
3. If all of the following conditions apply, set up a shell that does not reposition IP addresses before you execute the system reconfiguration command:
   The system is configured in a HiRDB parallel server configuration.
   The cluster software is HA monitor.
   The system switchover facility is in the server mode.
   The configuration is set up for inheriting IP addresses.
   

(a) Conditions under which the system reconfiguration command can be executed

When a recovery-unnecessary front-end server is used, the environments in which the system reconfiguration command can be executed differ. If any of the following conditions is satisfied, the system reconfiguration command cannot be executed:

• The pdstart -r command was used to start HiRDB.
• Some units or servers have been shut down (including reduced activation).^#
• A communication error has occurred between units in the network.
• Updatable online reorganization is being executed.
• The pdrplstop command is executing.
• The pd_mode_conf operand is set to AUTO.
• The front-end server in a stopped unit ceases to be a recovery-unnecessary front-end server after a system definition change.
• There is an in-memory RDAREA.
• The HiRDB expansion unit is being used (he pd_system_expand_unit operand is specified).

#: Not including a recovery-unnecessary front-end server.

Hint

The system reconfiguration command can be executed even if the recovery-unnecessary front-end server is not active. However, all units other than the recovery-unnecessary front-end server must be active. In addition, all servers in the active units must also be active.

(b) Note

Executing the system reconfiguration command also reconfigures the HiRDB system definition of units that are stopped (recovery-unnecessary front-end server units). During this process, stopped units are started in order to modify their HiRDB system definition.