8.8 Notes about cluster operation (for UNIX)
-
If you run multiple logical hosts concurrently in the cluster system, you need as many system resources as there are logical hosts running concurrently.
-
Before you set up JP1/IM - Manager in the cluster system, make sure that JP1/IM - Manager on the physical host has terminated. If you set up the cluster system while JP1/IM - Manager is running on the physical host, the logical host services will no longer function correctly. In such a case, restart the server to recover the system.
-
Before you start JP1/IM - Manager in a cluster system, make sure that you configure the authentication server that will be used on the logical host. For details about how to configure an authentication server, see the JP1/Base User's Guide. In addition, before you start JP1/IM - Manager, make sure that the configured authentication server is running.
-
When you set the user authentication server and register users on the logical host, make sure that you use the host at the primary node. Also make sure when you register users that you have already started the logical host services.
-
If server switching occurs at the user authentication server due to node switching during login processing, a communication failure occurs on JP1/IM - Manager. The error is recovered after the switching is completed.
If the problem is in the JP1/IM - Manager operation, you can avoid the problem by placing the user authentication server outside the cluster system.
-
Specify in the jbsgetcnf command used to back up the primary node definitions exactly the same case-sensitive logical host name that was specified when the logical host was defined.
If you specify the wrong name by mistake, you must delete the logical host and then specify the settings again.
-
If you do not use IM Configuration Management but distribute configuration definition information in the cluster system, create the configuration definition file under the following name:
shared-directory/jp1base/conf/route/jbs_route.conf
-
Do not rename hosts while JP1/IM - Manager is running by, for example, using a cluster software function.
If you have renamed hosts, see 2.2.5 Tasks to be performed before a logical host name is changed in a cluster system in the JP1/Integrated Management 2 - Manager Administration Guide and perform the required tasks.
-
If a logical host is created in a cluster system or the binding method is to be changed, be sure to stop the daemon of JP1/IM - Manager running on the physical host. If a logical host is created or the binding method is changed without stopping the daemon on the physical host, the daemon on the physical host will not operate normally. In this case, recover by rebooting the server machine.
-
If JP1/IM - Manager is not used on the physical host, be sure to cancel the setting of the automatic startup and stop script.
-
If you execute "jbsrt_get -h logical-host" on a standby host that uses a cluster, the message "KAVB3113-I Definition does not exist. " is displayed. If that happens, execute "jbsrt_get -h logical-host" on an active host.
-
For details on supported cluster software and notes on cluster software, see the following URL.