Job Management Partner 1/Performance Management User's Guide
This subsection describes the items to check before installing and setting up Performance Management in a cluster system.
- Organization of this subsection
- (1) Prerequisite conditions
- (2) Checking the setup environment
- (3) Notes about upgrading when a logical host is used
(1) Prerequisite conditions
The prerequisite conditions for using Performance Management in a cluster system are described below.
(a) Cluster system
Make sure that the following conditions are satisfied:
- The cluster system is controlled by cluster software.
- Settings are made so that the starting and stopping of Performance Management used on a logical host are controlled by cluster software.
- Note
- Suppressing dialog boxes when errors occur
- Because a failover might not be performed if an application error dialog box is displayed in Dr. Watson, you need to suppress the display of error dialog boxes. However, note that, if error notification is suppressed, information acquisition when an application error occurs might be affected.
- If an application error occurs in Windows, a dialog box for reporting the error to Microsoft sometimes appears. If this dialog box is displayed, a failover might not be performed, therefore, you need to suppress such error notification.
For details on how to do this, see the OS documentation.
(b) Shared disks
Make sure that the following conditions are satisfied:
- Each logical host has a shared disk, and the disk can be taken over by from the executing node by the standby node.
- Shared disks are physically connected to each node via Fibre Channel or SCSI. Configurations in which the shared disk is a network drive or disk replicated over a network are not supported.
- When a failover occurs, if some processes are still using the shared disks, make sure it is still possible to force shared disks offline via cluster software or by other means and perform a failover.
- If multiple Performance Management programs are executed on a single logical host, make sure the directory names for the shared disks are the same. For Store databases, make sure the storage destination can be changed to allow storage in a different directory on the same shared disk.
(c) Logical host name, logical IP address
Make sure that the following conditions are satisfied:
- There is a logical host name and corresponding logical IP address for each logical host, and switching from the executing node to the standby node can be performed.
- The logical host and logical IP address are set in the hosts file and on the name server.
- If using a DNS a logical host name is specified without a domain name, instead of by using a FQDN.
- Each physical host name and logical host name is unique within the system.
- Notes regarding logical host names:
- Do not use a physical host name (a host name displayed using the hostname command) for a logical host name. Otherwise, normal communication processing might be prevented.
- Logical host names must consist of from 1 to 32 bytes alphanumeric characters.
- localhost, an IP addresses, or a hyphen (-) cannot be used for a logical host name.
(2) Checking the setup environment
In addition to the environment information normally required to set up Performance Management, the following information is required to set up Performance Management used on a logical host.
Table 9-1 Information required to set up PFM - Manager to be used on a logical host (in Windows)
Item Example Logical host name jp1-ha1 Logical IP address 172.16.92.100 Shared disk S:\jp1 If multiple instances of Performance Management that use a single logical host, each instance uses the directory of the same shared disk.
(3) Notes about upgrading when a logical host is used
To upgrade PFM - Manager on a logical host, you must place a shared disk online on either an executing or a standby node.
However, there is no need to place a shared disk online to upgrade PFM - Web Console in a cluster environment.
All Rights Reserved. Copyright (C) 2009, Hitachi, Ltd.