Perform this procedure if Enterprise Manager clustering will be used in a two-tier system.
Procedure
Perform the procedure shown below.
3.8.1.1 Before constructing the environment
Set up the shared disk
Assign the IP address and host name
3.8.1.2 Installing on the server on the active node
Install an Enterprise Manager on the active server
Set up clustering on the active server
3.8.1.3 Installation on the standby node server
Install an Enterprise Manager on the standby server
Set up clustering on the standby server
Registering with the userApplication Configuration Wizard (Solaris versions)
Registering with the RMS Wizard (Linux versions)
This section explains how to prepare the following resources to build the cluster environment:
Shared disk used by the product
The shared disk is used to store files that are shared by the Enterprise Manager on the active node and the Enterprise Manager on the standby node.
Logical IP address used by this product
IP addresses that are unique within the network are newly assigned when operating the nodes in a cluster system.
Logical host name used by this product
This is a name that can be used to resolve the logical IP address.
It becomes valid when registered with a naming service such as the DNS.
Set up the shared disk
Set up a shared disk for storing files shared by the Enterprise Manager on the active node and the Enterprise Manager on the standby node. Note that the shared disk must be set up so that it is owned exclusively by Systemwalker Service Quality Coordinator.
If the storage location of the performance database (PDB) and the archive files are changed, make sure to configure the storage location physical disks as shared disks.
Refer to the PRIMECLUSTER manual for details on shared disk settings.
Assign the IP address and host name
When Systemwalker Service Quality Coordinator operates in an independent cluster system
Register the IP address and host name by referring to the PRIMECLUSTER manual.
When Systemwalker Service Quality Coordinator is added to a Systemwalker Centric Manager cluster system
For the logical IP address and the logical host name, use resources that have already been set up in the Systemwalker Centric Manager cluster system.
There is no need to create new resources.
Install an Enterprise Manager on the active server
Perform the installation process as described in "3.1.1 Installing a Manager/Enterprise Manager[EE]".
Note
Do not install an Enterprise Manager on the shared disk.
Set up clustering on the active server
Stop cluster applications on the active server
Use PRIMECLUSTER's Cluster Admin GUI to stop cluster applications.
Point
If Systemwalker Service Quality Coordinator is linked to Systemwalker Centric Manager, stop the Systemwalker Centric Manager service as well.
Enable access to the shared disk from the active server
Mount the shared disk on the active server.
Execute the cluster setup command on the active server
Execute the cluster setup command "sqcsetupclp" on the active server by referring to "A.6 Cluster Setup Commands".
The command format is as follows:
/opt/FJSVssqc/bin/sqcsetupclp -m <mount point> -h <logical host name> |
Change the storage location of the performance database (PDB) and the archive files
Refer to "6.6.1 Changing the PDB/Archive File Storage Location" for details.
Set up the Enterprise Manager
If the Agent functions included with the Enterprise Manager are used to collect performance information about the Enterprise Manager itself, create and apply a collection policy by referring to "A.1 Server Resource Information Collection Policy Creation Command".
Install an Enterprise Manager on the standby server
Install an Enterprise Manager by referring to "3.1.1 Installing a Manager/Enterprise Manager[EE]".
Note
Create the same Enterprise Manager environment (same device name, shared disk and installation path) on both the active and standby nodes.
Do not install an Enterprise Manager on the shared disk.
Set up clustering on the standby server
Stop cluster applications on the standby server
Use PRIMECLUSTER's Cluster Admin GUI to stop cluster applications.
Enable access to the shared disk from the standby server
Mount the shared disk on the standby server.
Execute the cluster setup command on the standby server
Execute the cluster setup command "sqcsetupclp" on the standby server by referring to "A.6 Cluster Setup Commands".
The command format is as follows:
/opt/FJSVssqc/bin/sqcsetupcls -m <mount point> |
Register the Enterprise Manager daemon as a PRIMECLUSTER resource.
The resource registration method for the active server is explained here.
Register the resource created using the Enterprise Manager cluster setup command as the Enterprise Manager daemon.
Point
If Systemwalker Service Quality Coordinator is synchronized with Systemwalker Centric Manager, register the Enterprise Manager daemon with the "CentricMGR Group" as an application resource.
The resource registration method differs according to the platform.
Refer to the registration method that matches the platform being used.
Registering with the userApplication Configuration Wizard (Solaris versions)
Create resources
Create cluster applications
Registering with the RMS Wizard (Linux versions)
Create resources
Create cluster applications
Registering with the userApplication Configuration Wizard (Solaris versions)
Create resources
Select "Create resource" in the PRIMECLUSTER userApplication Configuration Wizard and set the resource type. Refer to the PRIMECLUSTER manual for details.
Select a partition for shared resources.
Select "IP Address Takeover" as the network type and set the IP address or host name specified with the cluster setup command option as the IP address or host name.
Select "Application" as the procedure class and create a procedure resource as a SystemwalkerSQC resource.
Create cluster applications
Select "Create userApplication" from the top menu of the PRIMECLUSTER's userApplication Configuration Wizard to create the application. Refer to the PRIMECLUSTER manual for details.
Select "Standby" as the operation mode in the "Set userApplication Name and Operation Mode" window.
Set the userApplication attributes in the "Set Attributes" window. The following table shows the recommended values that can be specified with each attribute.
Attribute | Value | Overview |
|---|---|---|
AutoStartUp | Yes | userApplication will start automatically when RMS is started. |
AutoSwitchOver | HostFailure ResourceFailure | Failover will occur automatically when a problem occurs in a host or resource. |
PersistentFault | 1 | If the status of userApplication is "Faulted", it will continue even after RMS has restarted. |
ShutdownPriority | NONE (Default value) | A weight coefficient (priority) will not be assigned to userApplication. |
StandbyTransitions | StartUp SwitchRequest | A standby status change will occur when RMS starts up or when the status changes to a different node. |
OnlinePriority | 0 (Default value) | After RMS has restarted, userApplication will go online on the node with the highest priority. The priority is determined by the node setup sequence that is set with "SysNode Settings" when userApplication is created. |
HaltFlag | No (Default value) | If a fault occurs while a previous fault is being processed (i.e., a double fault exists), forced termination of the node is not performed. This will prevent userApplication from failing over. |
When selecting resources, select all the resources created in "1. Create resources" and then create applications.
Registering with the RMS Wizard (Linux versions)
This section explains how to use the PRIMECLUSTER RMS Wizard to create applications.
Create resources
Set logical IP address as a resource.
Select "Gls:Global-Link-Services" in the RMS Wizard, and set the logical IP address of PRIMECLUSTER GLS as a resource.
Set shared logical IP addresses as resources.
Creating a cluster application
Use the RMS Wizard to create an application.
Refer to the PRIMECLUSTER manual for details.
Primary/secondary node registration
Register primary and secondary nodes from "Machines+Basics".
Takeover IP address registration
Select "Gls:Global-Link-Services", and register the address to be used as the takeover IP address set with PRIMECLUSTER GLS.
Shared disk registration
From "LocalFileSystem", register the mount point of the partition for shared resources.
Set the Procedure resources
Set Procedure resources (SystemwalkerSQC).
Other optional settings
From "Machines+Basics", set the options for cluster applications. The following table shows some example option settings with recommended values:
Attribute | Value | Overview |
|---|---|---|
AutoStartUp | Yes | userApplication will start automatically when RMS is started |
AutoSwitchOver | HostFailure ResourceFailure | Failover will occur automatically when a problem occurs in a host or resource. |
PersistentFault | 1 | If the status of userApplication is "Faulted", it will continue even after RMS has restarted. |
ShutdownPriority | NONE (Default value) | A weight coefficient (priority) will not be assigned to userApplication. |
StandbyTransitions | StartUp SwitchRequest | A standby status change will occur when RMS starts up or when the status changes to a different node. |
OnlinePriority | 0 (Default value) | After RMS has restarted, userApplication will go online on the node with the highest priority. The priority is determined by the node setup sequence that is set with "SysNode Settings" when userApplication is created. |
HaltFlag | No (Default value) | If a fault occurs while a previous fault is being processed (i.e., a double fault exists), forced termination of the node is not performed. This will prevent userApplication from failing over. |
Use PRIMECLUSTER's Cluster Admin GUI to start the Systemwalker Service Quality Coordinator daemon.
The environment settings of the PRIMECLUSTER cluster system are now complete.
The next step is to install and set up the environment according to the operation model that will be adopted.
Note
If a two-tier operation model environment for a Manager is being constructed, installation of the Enterprise Manager is now complete and the procedure described in "Tasks to be performed on an Enterprise Manager" can be skipped.
