This subsection describes the automatic restart of logical servers, when a failure occurs.

When a failure occurs, a stopped logical server can restart automatically using the Management Server.

Automatic restart is executed when the status after startup is detected as 'failure', from the logical server that was started normally by the Management Server. 'Failure' is a state when the logical server stops without receiving any stop request.

Administration Agent monitors the processes of the logical server and checks the operations of the logical server. When a logical server failure, such as a process is down or has hung up is found, Administration Agent detects the failure and notifies Management Server. When a failure is detected, the Management Server executes the commands and collects the snapshot logs, after that the Management Server collects the data used for troubleshooting, and then the logical server restarts automatically.

When a failure occurs during the startup process requested by the user, instead of automatically restarting the logical server, the user is notified of the failure to start the logical server.

Automatic restart is executed according to the automatic restart frequency and automatic restart retry interval that is specified when building the system. Note that automatic restart is not executed when the specified automatic restart frequency is '0'.

When automatic restart exceeds the specified frequency, a message is output to Manager-log-output-directory/mngsvr-number-of-files.log and the logical server status changes to 'failure'.

If true is specified in the com.cosminexus.mngsvr.logical_server_abnormal_stop.exit key of the mserver.properties file (Management Server environment configuration file), Management Server stops when the logical server status changes to 'abnormal termination'. 'Abnormal termination' is a state wherein the logical server stops when the automatic restart frequency is exceeded or when 0 is specified as the automatic restart frequency and a failure is detected.

The following figure shows the flow of automatic restart when a failure occurs and the status of the logical server:

Figure 2-3 Automatic restart when failure occurs

Note that when an automatic restart is executed in the requisite logical server, then after the requisite logical server has automatically restarted, the logical server for which the corresponding logical server is specified as a pre-requisite is also restarted.

Hint

Detection of failure

Check the status of the logical server by monitoring the process of the logical server, and checking the operation of the logical server in the Administration Agent. For example, process monitoring checks the existence of the process ID of the logical server process. If the process ID does not exist, the Administration Agent detects that the process is down and notifies the error to the Management Server.

The contents of process monitoring and checking the operations differ depending upon the type of the logical server. For details, see 2.3 Mechanism of Starting and Stopping the Logical Server.

If the logical server process ends immediately after successfully starting the logical server, the error detection and automatic restart might be executed again. In such cases, you can avoid this process by setting a value more than the time from when the process of the logical server starts until the process ends in the adminagent.server-type.watch.start_time key (the time from executing the start command of the logical server until the starting of operation confirmation) of adminagent.properties.