ETERNUS SF Disk Space Monitor User's Guide - Microsoft(R) Windows Server(TM) 2003/Solaris(TM) Operating System/Linux - 13.0

4.4 Cluster System Operation Model

Operating Managers in a cluster system improves the availability of management operations. This is because even if a fault occurs with one node in the cluster system, management operations can be taken over by the other node, which is still functioning correctly.

Managers cannot be installed on cluster systems in the following situations.

• Machines where both the Manager and the operation management client have been installed
• Managers that use redundant operations

The following cluster systems and configurations are supported.

[Cluster system]

Microsoft(R) Cluster Server (MSCS) included in the following products:

• Microsoft(R) Windows Server(TM) 2003, Enterprise Edition
• Microsoft(R) Windows Server(TM) 2003, Datacenter Edition

• Fujitsu PRIMECLUSTER Enterprise Edition 4.1
• Fujitsu PRIMECLUSTER HA Server 4.1
• Fujitsu PRIMECLUSTER Clustering Base 4.1

(Linux)

• Fujitsu PRIMECLUSTER Enterprise Edition 4.1A20
• Fujitsu PRIMECLUSTER HA Server 4.1A20
• Fujitsu PRIMECLUSTER Clustering Base 4.1A20

(Linux for Itanium)

• Fujitsu PRIMECLUSTER Enterprise Edition 4.1A40
• Fujitsu PRIMECLUSTER HA Server 4.1A40

[Cluster configuration]

• 1:1 active/standby node configuration (configurations where one node is on standby for another node)

In this section, the node that performs management tasks is referred to as the "active node", while the node that is standing by to take over these tasks is referred to as the "standby node". The operation for transferring management tasks from the active node to the standby node when an error occurs is called "failover". Also, the disk that is shared by the active node and the standby node is referred to as the "shared disk".

Use logical host names or logical IP addresses to specify the host names or IP addresses for Managers in a cluster system (such as when specifying the connection destination Manger from the Console or during the installation of an Agent).

This allows ETERNUS SF Disk Space Monitor to be operated without the need to be aware of which node is currently active.

The installation procedure differs depending on the type of cluster system being installed. Refer to the installation method corresponding to the cluster system being installed.

Cluster system	Reference location
MSCS	Section 4.4.1, "Operating MSCS cluster systems"
PRIMECLUSTER	Section 4.4.2, "Operating PRIMECLUSTER cluster systems"

4.4.1 Operating MSCS cluster systems

4.4.1.1 Installation and environment setup 

This section explains how to install ETERNUS SF Disk Space Monitor on MSCS cluster systems. Be sure to follow the order shown.

• This section explains the procedure for creating a ETERNUS SF Disk Space Monitor environment in an MSCS cluster environment and so includes MSCS setup operations. Reading this section requires a basic understanding and some practical knowledge of MSCS.
• Use an account with Administrator privileges to perform the settings in this section.

4.4.1.1.1 Creating a group and registering basic resources 

MSCS must be installed and the MSCS environment must be set up before the operations explained in this section can be performed.

If necessary, refer to the MSCS manual for the procedures for installing MSCS and creating a new cluster environment.

Creating a group

Register the group to be used by ETERNUS SF Disk Space Monitor.

• Use the cluster administrator to create a resource group named "DiskSpaceMonitor Group" for this product.

If this product is linked with Systemwalker Centric Manager, a resource group need not be created.

Setting up the shared disk, host name and IP address

Register the basic resources with the group.

• Register host names, IP addresses, and the disk for storing shared files for this product with the "DiskSpaceMonitor Group".

If this product is linked with Systemwalker Centric Manager, register the above resources in the "CentricMGR Group".

4.4.1.1.2 Installing ETERNUS SF Disk Space Monitor on the active server 

1. Install ETERNUS SF Disk Space Monitor on the active server.

   

   Do not install ETERNUS SF Disk Space Monitor on the shared disk at this stage.

   

   Refer to Section 4.1.1.1, "Installation", for details on the installation method.

4.4.1.1.3 Executing the cluster setup command on the active server 

1. Vary the shared disk online.
2. Get disk ownership rights on the active server.
   Get the following disk ownership rights:
   • Quorum disk
   • The shared disk used by this product
3. Close the cluster administrator and the registry editor.
   If the cluster administrator and the registry editor are running, close them.
4. Execute the cluster setup command on the active server.
   Execute the sqcsetupclp cluster setup command on the active server. The syntax for this command is as follows:

   Installation directory\bin\sqcsetupclp -m shared disk -h logical host name

   

   Refer to Section 4.4.3, "Cluster setup command" for details on the cluster setup command.

4.4.1.1.4 Installing ETERNUS SF Disk Space Monitor on the standby server 

1. Install ETERNUS SF Disk Space Monitor on the standby server.

   

   When installing ETERNUS SF Disk Space Monitor, make sure both the active server and the standby server have the same environment (the same directory configuration).

   Also, do not install ETERNUS SF Disk Space Monitor on the shared disk.

   

   Refer to Section 4.1.1.1, "Installation" for details on the installation method.

4.4.1.1.5 Executing the cluster setup command on the standby server 

1. Close the cluster administrator and the registry editor.
   If the cluster administrator and the registry editor are running, close them.
2. Execute the cluster setup command on the standby server.
   Execute the sqcsetupcls cluster setup command on the standby server. The syntax for this command is as follows:

   Installation directory\bin\sqcsetupcls -m shared disk

   

   Refer to Section 4.4.3, "Cluster setup command" for details on the cluster setup command.

4.4.1.1.6 Registering resources

Register the ETERNUS SF Disk Space Monitor services as MSCS resources on the active server.

Register the following services as MSCS application resources in the "DiskSpaceMonitor Group" created in Creating a group.

• ETERNUS SF Disk Space Monitor DCM service
• ETERNUS SF Disk Space Monitor sqcschdle service

If this product is linked with Systemwalker Centric Manager, register all of these services as resources in the "CentricMGR Group".

ETERNUS SF Disk Space Monitor DCM service

Item	Setting
Name	ETERNUS SF Disk Space Monitor DCM
Resource type	Generic service
Group	DiskSpaceMonitor Group or CentricMGR Group
Owners that can execute this service	Active server and standby server
Resource dependencies	The IP address and the shared disk
Service name	SQC_DCM

ETERNUS SF Disk Space Monitor sqcschdle service

Item	Setting
Name	ETERNUS SF Disk Space Monitor sqcschdle
Resource type	Generic service
Group	DiskSpaceMonitor Group or CentricMGR Group
Owners that can execute this service	Active server and standby server
Resource dependencies	The IP address and the shared disk
Service name	sqcschdle

4.4.1.2 Uninstallation 

Use the following procedure to remove ETERNUS SF Disk Space Monitor from cluster systems.

1. Make the current server the active node.
2. Stop the services that are being managed by the cluster system.
   From the cluster administrator, vary the following ETERNUS SF Disk Space Monitor services offline.
   • ETERNUS SF Disk Space Monitor DCM service

   

   If this product is linked with Systemwalker Centric Manager, stop the Systemwalker Centric Manager service as well.

3. Remove resources.
   Use the cluster administrator to remove the application resources and ETERNUS SF Disk Space Monitor cluster services registered in Section 4.4.1.1.6, "Registering resources".
4. Execute the cluster cancel command on the standby server.
   The syntax for this command is as follows:

   Installation directory\bin\sqcunsetcl

5. Uninstallation
   Uninstall ETERNUS SF Disk Space Monitor from the standby server.

   

   If necessary, refer to Section 4.1.4, "Uninstallation" for details on uninstallation.

6. Execute the cluster cancel command on the active server.
   The syntax for this command is as follows:

   Installation directory\bin\sqcunsetcl

7. Uninstallation
   Uninstall ETERNUS SF Disk Space Monitor from the active server.

   

   If necessary, Section 4.1.4, "Uninstallation" for details on uninstallation.

8. Delete unnecessary directories from the shared disk.
9. Delete the shared disk.
   Use the cluster administrator to delete the shared disk (that has been registered as a cluster resource and has been used by ETERNUS SF Disk Space Monitor for cluster operations) from the "DiskSpaceMonitor Group".

   

   If this product is linked with Systemwalker Centric Manager, delete this shared disk from the "CentricMGR Group" as well.

10. Delete the "DiskSpaceMonitor Group".

    

    If this product is linked with Systemwalker Centric Manager, there is no need to delete the group.
    Start the Systemwalker Centric Manager service that was stopped in Step 2.

4.4.2 Operating PRIMECLUSTER cluster systems

4.4.2.1 Installation and environment setup 

This section explains how to install ETERNUS SF Disk Space Monitor on PRIMECLUSTER cluster systems. Perform these installation and setup operations in the same order as they are presented.

• This section explains the procedure for creating a ETERNUS SF Disk Space Monitor environment in a PRIMECLUSTER cluster environment and so includes PRIMECLUSTER setup operations. Reading this section requires a basic understanding and some practical knowledge of PRIMECLUSTER.
• Log in with system administrator privileges to make the settings described in this section.

4.4.2.1.1 Setting up the shared disk and allocating IP addresses and host names 

PRIMECLUSTER must be installed and the PRIMECLUSTER environment must be set up before the operations explained in this section can be performed.

If necessary, refer to the PRIMECLUSTER manual for the procedures for installing PRIMECLUSTER and creating a new cluster environment.

Setting up the shared disk

Set up the shared disk that stores files that are shared by ETERNUS SF Disk Space Monitor on the active and standby nodes.

Refer to the PRIMECLUSTER manual for details on how to set up the shared disk device.

Allocating the IP addresses and host names

If this product is linked with Systemwalker Centric Manager, there is no need to allocate a new IP address and host name.

Each node in the cluster system manages multiple IP addresses and host names.

• The IP address of the group used by ETERNUS SF Disk Space Monitor (hereafter referred to as the "logical IP address")
  When the cluster system starts operating, the network allocates a new, unique IP address.
• The host name for ETERNUS SF Disk Space Monitor (hereafter referred to as the "logical host name")
  This host name resolves to the logical IP address.
  This host name can be made valid by registering it with a naming service such as DNS.

Refer to the PRIMECLUSTER manual for details on how to set IP addresses and host names.

4.4.2.1.2 Installing ETERNUS SF Disk Space Monitor on the active server 

1. Install ETERNUS SF Disk Space Monitor on the active server.

   

   Do not install ETERNUS SF Disk Space Monitor on the shared disk at this stage.

   

   Refer to Section 4.1.1.1, "Installation" for details on the installation method.

4.4.2.1.3 Executing the cluster setup command on the active server 

1. Stop the cluster application on the active server.

   

   If this product is linked with Systemwalker Centric Manager, stop the Systemwalker Centric Manager service as well.

2. Make the shared disk device accessible from the active server.
3. Execute the cluster setup command on the active server.
   Execute the sqcsetupclp cluster setup command on the active server. The syntax for this command is as follows:

   /opt/FJSVssqc/bin/sqcsetupclp -m mount point for the shared disk device -h logical host name

   
   

   

   Refer to Section 4.4.3, "Cluster setup command" for details on the cluster setup command.

4.4.2.1.4 Installing ETERNUS SF Disk Space Monitor on the standby server 

1. Install ETERNUS SF Disk Space Monitor on the standby server.

   

   Refer to Section 4.1.1.1, "Installation" for details on the installation method.

   

   Create the same ETERNUS SF Disk Space Monitor environment on both the active server and the standby server (the same device name, shared disk device, and installation path).
   Do not install ETERNUS SF Disk Space Monitor on the shared disk at this stage.

4.4.2.1.5 Executing the cluster setup command on the standby server 

1. Stop the cluster application on the standby server.
2. Make the shared disk device accessible from the standby server.
3. Execute the cluster setup command on the standby server.
   Execute the sqcsetupcls cluster setup command on the standby server. The syntax for this command is as follows:

   /opt/FJSVssqc/bin/sqcsetupcls -m mount point for the shared disk device

   

   Refer to Section 4.4.3, "Cluster setup command" for details on the cluster setup command.

4.4.2.1.6 Registering resources

The ETERNUS SF Disk Space Monitor daemon is registered as a PRIMECLUSTER resource

This section explains how to register the resource on the active server

The overall procedure is as follows:

• Register the ETERNUS SF Disk Space Monitor resource created by the cluster setup command as the ETERNUS SF Disk Space Monitor daemon.

  

  If this product is linked with Systemwalker Centric Manager, register the ETERNUS SF Disk Space Monitor daemon as an application resource in the 'CentricMGR Group'.

The method for registering resources is different for each platform.

Refer to the registration method for the platform being used.

Platform	Reference location
Solaris	Section 4.4.2.1.7, "Registering resources using the userApplication Configuration Wizard"
Linux	Section 4.4.2.1.8, "Registering resources using the RMS Wizard"

4.4.2.1.7 Registering resources using the userApplication Configuration Wizard

Creating resources

From PRIMECLUSTER's userApplication Configuration Wizard, select Create Resource, and set up the following Resource types. Refer to the PRIMECLUSTER manual for details.

• Fsystem:

  Select a partition for shared resources.

• Ipaddress:

  Select IP Address Takeover for the network type. For the IP address/host name, use the one specified in the cluster setup command option.

• Procedure:

  Select Application for the procedure class and DiskSpaceMonitor for the procedure resource.

Creating cluster applications

Create the application by selecting Create userApplication from the top menu of PRIMECLUSTER's userApplication Configuration Wizard. Refer to the PRIMECLUSTER manual for details.

• In the userApplication name and operation type setting window, select Standby for the operation type.
• In the Set Attributes window, set the userApplication attributes. The following table shows the example settings (recommended values) of these attributes.

  Attribute	Value	Outline
  AutoStartUp	Yes	When RMS starts, the userApplication also automatically starts.
  AutoSwitchOver	HostFailure ResourceFailure	This enables automatic failover when host or resource fails.
  PersistentFault	1	When userApplication is in Faulted state, it stays in this state even after RMS reboot.
  ShutdownPriority	NONE (default)	A weighting factor (priority) is not allocated to the userApplication.
  StandbyTransitions	StartUp SwitchRequest	Transition to standby state occurs at RMS startup or on transition to another node.
  OnlinePriority	0 (default)	After the RMS restarts, the userApplication goes online at the node with the highest priority. The priority is determined by the node setup order that was specified in 'SysNode setting' at userApplication creation time.
  HaltFlag	No (default)	If another problem (double failure) occurs during Fault processing, forced shutdown of the node does not occur. As a result, the userApplication will not failover.

• Select all the resources created in "Creating resources" and then create the application.

4.4.2.1.8 Registering resources using the RMS Wizard

This section explains the settings required when creating an application using PRIMECLUSTER's RMS Wizard.

Preliminary settings

Settings are necessary in advance because logical IP addresses and shared disks will be used.

Refer to the PRIMECLUSTER manual for how to make settings.

Creating the cluster application

Start the RMS Wizard and register DiskSpaceMonitor as an application resource.

Specify a logical IP address in IpAddress and the shared disk mount point in LocalFileSystem.

4.4.2.2 Uninstallation 

This section describes the procedure for removing ETERNUS SF Disk Space Monitor from cluster systems.

1. Stop the cluster application.
   Use the PRIMECLUSTER Cluster Admin GUI to stop the ETERNUS SF Disk Space Monitor service.

   

   Refer to the PRIMECLUSTER manual for details on how to stop services.

   

   If this product is linked with Systemwalker Centric Manager, stop the Systemwalker Centric Manager service as well.

2. Delete the application and its resources
   Delete the ETERNUS SF Disk Space Monitor application and its resources that were registered with the cluster system in Section 4.4.2.1.6, "Registering resources".
3. Execute the cluster cancel command on the standby server.
   Vary the shared disk online, and execute the sqcunsetcl cluster cancel command on the standby server. The syntax for this command is as follows:

   /opt/FJSVssqc/bin/sqcunsetcl

4. Execute the cluster cancel command on the active server.
   Execute the sqcunsetcl cluster cancel command on the active server. The syntax for this command is as follows:

   /opt/FJSVssqc/bin/sqcunsetcl

5. Uninstall ETERNUS SF Disk Space Monitor.
   Uninstall ETERNUS SF Disk Space Monitor on both the active server and the standby server.

   

   If necessary, refer to Section 4.1.4, "Uninstallation" for details on uninstallation.

6. Delete unnecessary directories from the shared disk device.
7. Delete the shared disk device used by ETERNUS SF Disk Space Monitor from the ETERNUS SF Disk Space Monitor cluster application.
   Delete the following shared disk device that was registered in "Setting up the shared disk".
   • Delete the shared disk device used by ETERNUS SF Disk Space Monitor from the ETERNUS SF Disk Space Monitor cluster application.

   

   If this product is linked with Systemwalker Centric Manager, delete this shared disk from the "CentricMGR Group" as well.

8. Delete the ETERNUS SF Disk Space Monitor cluster application.

   

   If this product is linked with Systemwalker Centric Manager, there is no need to delete the cluster application.

   Start the Systemwalker Centric Manager cluster application that was stopped in step 1.

4.4.3 Cluster setup command

The cluster setup command must be executed to create the cluster environment.
The specification for the cluster setup command is explained below.

This command can be executed on Managers.

[Command name]

sqcsetupclp: Cluster environment creation command (active node)

sqcsetupcls:Cluster environment creation command (standby node)

[Syntax]

(Active node)

Installation directory\bin\sqcsetupclp -m shared disk -h logical host name

(Standby node)

Installation directory\bin\sqcsetupcls -m shared disk

(Active node)

/opt/FJSVssqc/bin/>/bin/sqcsetupclp -m mount point -h logical host name

(Standby node)

/opt/FJSVssqc/bin/>/bin/sqcsetupcls -m mount point

[Function]

This command creates a cluster environment for ETERNUS SF Disk Space Monitor.

[Options]

Option	Meaning
-m shared disk	Specifies the drive for the shared disk device to be used by the cluster operation.
-h logical host name	Specifies the logical host name or logical IP address to be used by the cluster operation.

Option	Meaning
-m mount point	Specifies the mount point for the shared disk device to be used by the cluster operation.
-h logical host name	Specifies the logical host name or logical IP address to be used by the cluster operation.

[Termination status]

Normal termination: 0

Abnormal termination: 1

If this command terminates normally, the following message will be output:

• "Cluster setup succeeded"

[Usage example]

(Active node: sqcsetupclp)

C:\>cd C:\Program Files\DiskSpaceMonitor\bin

C:\Program Files\DiskSpaceMonitor\bin>sqcsetupclp -m F:\ -h hostname

126 files copied.

0 files copied.

0 files copied.

0 files copied.

0 files copied.

0 files copied.

0 files copied.

0 files copied.

0 files copied.

0 files copied.

0 files copied.

0 files copied.

0 files copied.

Cluster setup succeeded

C:\Program Files\DiskSpaceMonitor\bin>

(Standby node: sqcsetupcls)

C:\>cd C:\Program Files\DiskSpaceMonitor\bin

C:\Program Files\DiskSpaceMonitor\bin>sqcsetupcls.exe -m F:\

Cluster setup succeeded

C:\Program Files\DiskSpaceMonitor\bin>

(Active node: sqcsetupclp)

# cd /opt/FJSVssqc/bin/

# ./sqcsetupclp -m /share -h hostname

Cluster setup succeeded

#

(Standby node: sqcsetupcls)

# cd /opt/FJSVssqc/bin/

# ./sqcsetupcls -m /share

Cluster setup succeeded

#

[Notes]

Execute this command with Administrator privileges.

Execute this command with superuser (root) privileges.