17.10.4 Starting and stopping the system during maintenance of the 1-to-1 node switching system of Application Server

This subsection describes the procedure for starting and stopping the system when you perform maintenance of the 1-to-1 node switching system of Application Server.

The stopping procedure for the following three cases is as follows:

When you do not need to restart the system during maintenance
When you need to restart the system during maintenance (stopping both nodes concurrently)
When you need to restart the system during maintenance (avoid stopping both the nodes concurrently)

Note that these procedures are applicable when Application Server is already running in the executing node host.

Important note: If node switching occurs as a result of failure of the Administration Agent, the processes of each logical server remains in the host in which the Administration Agent has failed. If processes are still running, you cannot start the logical servers when node switching occurs, and therefore, you must stop all processes before occurrence of node switching. Apart from this, execute the tasks for proper start of applications beforehand.

Organization of this subsection

(1) When you do not need to restart the system during maintenance
(2) When you need to restart the system during maintenance (How to stop both the nodes concurrently)
(3) When you need to restart the system during maintenance (How to avoid stopping both the nodes concurrently)

(1) When you do not need to restart the system during maintenance

To start and stop the system when you do not need to perform restart during maintenance:

In the executing node host, stop the running J2EE application and resource adapter.

In case of a batch server, check that the batch application is not running, and then stop the resource adapter.

For stopping the system, see the following manuals:
- In a system executing J2EE applications
  
  See 4.1.3 Procedure for stopping a system and 4.1.4 Methods of stopping a system in the uCosminexus Application Server Management Portal User Guide.
- In a system executing batch applications
  
  See 6.1.3 Procedure for stopping a system and 6.1.4 Methods of stopping a system in the uCosminexus Application Server Management Portal User Guide.
Use the server management commands to execute the maintenance process.

For operations of the server management commands, see 3. Basic Operations of Server Management Commands in the uCosminexus Application Server Application Setup Guide.
Start the J2EE application and resource adapter that you stopped in step 1.

In case of the J2EE server, start the J2EE application and resource adapter. In case of the batch server, start the resource adapter.

For the startup methods, see the following manuals:
- In a system executing J2EE applications
  
  See 4.1.1 Procedure for starting a system and 4.1.2 Methods of starting a system in the uCosminexus Application Server Management Portal User Guide.
- In a system executing batch applications
  
  See 6.1.1 Procedure for starting a system and 6.1.2 Methods of starting a system in the uCosminexus Application Server Management Portal User Guide.
As and when required, check the operation of the J2EE application and resource adapter.
In the executing node host (active node host), execute the monswap command and switch the nodes.

The executing node host (active node host) changes to the standby node host, and the host that earlier was the standby node (spare-node host) becomes the executing node host.
In the executing node host following switching (spare-node host), execute step 1. to step 4.
In the post-switching executing node host (spare-node host), execute the monswap command again as and when required.

Execute this step to return the active node host to the executing node.

This completes the start and stop process during maintenance.

To Page Top

(2) When you need to restart the system during maintenance (How to stop both the nodes concurrently)

Following is the procedure to start and stop the system by stopping both the nodes concurrently when you need to perform restart during maintenance:

When you have not specified the settings for batch stop of the logical servers of the management domain in the script for stopping the Administration Agent, stop each logical server in the executing node host.

This step is not required when you have specified the settings for stopping the logical servers in the script for stopping the Administration Agent. For stopping the logical servers, see the following manuals:
- In a system executing J2EE applications
  
  See 4.1.4 Methods of stopping a system in the uCosminexus Application Server Management Portal User Guide.
- In a system executing batch applications
  
  See 6.1.4 Methods of stopping a system in the uCosminexus Application Server Management Portal User Guide.
In the executing node host, execute the monend command and stop the system.
```
# monend server-identification-name
```
In the underlined part, specify the identification name of the server specified in the operand alias in the servers file.

The executing node host stops. Also, an instruction for stopping is output automatically in the standby node by the HA monitor, and the standby node host also stops.
In the active node and spare node host, execute the maintenance operation including changing the definition file.
In the active node host, execute the monbegin command and start the Administration Agent of the active node host.
```
# monbegin server-identification-name
```
In the underlined part, specify the identification name of the server specified in the operand alias in the servers file.

By doing this, the active node host starts operating as the executing node.
Execute the monbegin command in the spare-node host.
```
# monbegin server-identification-name
```
In the underlined part, specify the identification name of the server specified in the operand alias in the servers file.

By doing this, the spare-node host starts operating as the standby node.
When you have not specified the settings for batch start of the logical servers of management domain in the script for starting the Administration Agent, start each logical server of the executing node host.

This step is not required when you have specified the settings for starting the logical servers in the script for starting the Administration Agent. For starting the logical servers, see the following manuals:
- In a system executing J2EE applications
  
  See 4.1.2 Methods of starting a system in the uCosminexus Application Server Management Portal User Guide.
- In a system executing batch applications
  
  See 6.1.2 Methods of starting a system in the uCosminexus Application Server Management Portal User Guide.
In the executing node host, confirm the operation related to definition changes as and when required.

Reference note: For details on defining the servers file (setting the server environment with the HA monitor), see 17.6.6 Setting the server-compliant environment.

To Page Top

(3) When you need to restart the system during maintenance (How to avoid stopping both the nodes concurrently)

To start and stop the system by stopping both nodes concurrently when you need to perform restart during maintenance:

In the standby node host, execute the maintenance operation including changing the definition file.
In the executing node host, execute the monswap command and switch the nodes.
```
# monswap server-identification-name
```
In the underlined part, specify the identification name of the server specified in the operand alias in the servers file.

Application Server running on the executing node host stops, and Application Server starts on the node that was in the standby status until now. As a result, the host for which you have already performed maintenance becomes the executing node host.
In the post-switching executing node host (spare node host), confirm the operation related to definition changes as and when required.
In the post-switching standby node host (active node host), execute the maintenance operation including editing of definition files.
In the post-switching executing node host (spare node host), execute the monswap command again as and when required.

Execute this step when returning the active node host to the executing node.
In the post-switching executing node host (active node host), confirm the operation related to definition changes as and when required.

This completes the start and stop process during maintenance.

Reference note: For details on defining the servers file (setting the server environment with the HA monitor), see 17.6.6 Setting the server-compliant environment.

To Page Top