PRIMECLUSTER Installation and Administration Guide 4.2 (Linux for Itanium)

9.2 Maintenance Flow

Your Fujitsu customer engineer (CE) shall perform parts replacement. The flow of maintenance is as follows:

When stop maintenance is to be performed

1. All nodes of the running PRIMECLUSTER system shall be stopped by the administrator of the PRIMECLUSTER system.

2. Pass the operation over to your Fujitsu customer engineer (CE).

3. Your Fujitsu customer engineer (CE) shall then perform maintenance of the erroneous location (repair or replacement). Confirm that the system operates normally by running a test program, etc.

4. After the completion of maintenance by your Fujitsu customer engineer (CE), check the relevant equipment and then boot the PRIMECLUSTER system.

When job hot maintenance is to be performed

1. The administrator of the PRIMECLUSTER system shall shut down the node that contains the target equipment, so as to separate it from the operation, and then pass the operation over to your Fujitsu customer engineer (CE).

   For details on how to separate the node from the operation, see "Detaching Resources from Operation."

2. Your Fujitsu customer engineer (CE) shall confirm the target equipment and perform maintenance of the erroneous equipment (repair or replacement). Operation shall be confirmed by using a test program, etc.

3. After your Fujitsu customer engineer (CE) completes the maintenance and confirms the operation of the relevant equipment, boot the node and then execute standby restoration for the operation.

   For details on standby restoration for the operation, see "Executing Standby Restoration for an Operating Job."

---

9.2.1 Detaching Resources from Operation

9.2.2 Executing Standby Restoration for an Operating Job