2.7 Settings for Automatic Reflection

Perform settings so that the following information can be automatically reflected on the active and standby nodes:

Security information
Calendar information
- Day change time
- Calendar holiday information
- Schedule information, power schedule information and completion monitoring information for SYSTEM_CALENDAR
Service/application startup information
- Application startup information

These settings are not required on the execution server that configures the cluster system.

Settings for automatically reflecting security information

This section describes the procedure for automatically reflecting security information.

Note

If many users should be given the access right, it takes time to switch clusters.

As a guide, if the access right is given to more than 100 users, confirm the time required to switch clusters by performing a verification.

If the switch takes more time than planned, consider one of the following measures.

Create a group for the users with the access right.
Delete unnecessary users.
[Solaris] [Linux] Review (extend) the monitoring start time (60 seconds by default) in the monitoring script
The following sample monitoring script is provided:
- Solaris: /opt/FJSVJMCMN/etc/script/omgr_smonitor
- Linux: /opt/FJSVJMCMN/etc/script/omgr_smonitor

Settings for 1:1 active/standby (without subsystems), 1:1 active/standby (with subsystems), N:1 active/standby, and cascading configurations

For PRIMECLUSTER (for Solaris or Linux version), Oracle Solaris Cluster

Use the following procedure to perform settings on the active node. The command line provided here as a sample is for Solaris version.

Mount the shared disk. In the following example, the mount point of the shared disk is "/disk1".
```
# mount /disk1 (*1)
```
*1
The command line for Linux version is as follows. (In this example, the device is "/dev/sdb1" and the shared disk is "/disk1".)
```
# /bin/mount /dev/sdb1 /disk1
```

For N:1 active/standby configuration, create symbolic links to the shared disk.

# rm /var/opt/FJSVfwseo/JM 2> /dev/null
# ln -s /disk1/FJSVfwseo/JM /var/opt/FJSVfwseo/JM

Set up cluster information by executing the following command:
```
# mpaclcls
```
Refer to the Systemwalker Operation Manager Reference Guide for details on the mpaclcls command.
Reflect security information in files on the shared disk by executing the following command:
```
# mpcssave
```
Refer to the Systemwalker Operation Manager Reference Guide for details on the mpcssave command.
Unmount the shared disk.
```
# umount /disk1 (*1)
```
*1
The command line for Linux version is as follows.
```
# /bin/umount /disk1
```

For N:1 active/standby configuration, perform steps 1 to 5 on each active node.

In this case, specify the shared disk of each active node.

For MC/ServiceGuard

Use the following procedure to perform settings on the active node.

Mount the shared disk. In the following example, the device is "/dev/vg01" and the mount point of the shared disk is "/disk1".
```
# vgchange -a e /dev/vg01
# mount /disk1
```

For N:1 active/standby configuration, create symbolic links to the shared disk.

# rm /var/opt/FJSVfwseo/JM 2> /dev/null
# ln -s /disk1/FJSVfwseo/JM /var/opt/FJSVfwseo/JM

Set up cluster information by executing the following command:
```
# mpaclcls
```
Refer to the Systemwalker Operation Manager Reference Guide for details on the mpaclcls command.
Reflect security information in files on the shared disk by executing the following command:
```
# mpcssave
```
Refer to the Systemwalker Operation Manager Reference Guide for details on the mpcssave command.

Unmount the shared disk.

# umount /disk1
# vgchange -a n /dev/vg01

For N:1 active/standby configuration, perform steps 1 to 5 on each active node.

In this case, specify the shared disk of each active node.

For HACMP

Use the following procedure to perform settings on the active node.

Mount the shared disk. In the following example, the volume name of the shared disk is "datavg1" and the mount point to the shared disk is "/disk1".
```
# varyonvg datavg1
# mount /disk1
```
Set up cluster information by executing the following command:
```
# mpaclcls
```
Refer to the Systemwalker Operation Manager Reference Guide for details on the mpaclcls command.
Reflect security information in files on the shared disk by executing the following command:
```
# mpcssave
```
Refer to the Systemwalker Operation Manager Reference Guide for details on the mpcssave command.
Unmount the shared disk.
```
# umount /disk1
# varyoffvg datavg1
```

Settings for 1:1 active/standby configuration (with subsystems and partial cluster operation)

Note

These settings should only be applied to subsystems that will be registered with a cluster system.

For PRIMECLUSTER (for Solaris or Linux version)

This section explains the setup procedure using an example where Subsystem 1 is used in cluster operation and "/disk1" is designated as the mount point of the shared disk. The command line provided here as a sample is for the Solaris version.

Perform the following settings at the active node.

Mount the shared disk.
```
# mount /disk1(*1)
```
*1
The command line for Linux version is as follows. (In this example, the device is "/dev/sdb1" and the shared disk is "/disk1".)
```
# /bin/mount /dev/sdb1 /disk1
```
Execute the following command to set the cluster information:
```
# mpaclcls-s 1
```
Refer to the Systemwalker Operation Manager Reference Guide for details on the mpaclcls command.
Execute the following command to apply the security information to files on the shared disk:
```
# mpcssave-s 1
```
Refer to the Systemwalker Operation Manager Reference Guide for details on the mpcssave command.
Unmount the shared disk.
```
# umount /disk1(*1)
```
*1
The command line for Linux version is as follows:
```
# /bin/umount /disk1
```

For MC/ServiceGuard

In the examples that explain the migration procedure shown below, Subsystem 1 runs as a cluster system, the device is "/dev/vg01", and the mount point to the shared disk is "/disk1".

Use the following procedure to configure the settings on the active node.

Mount the shared disk.

# vgchange -a e /dev/vg01
# mount /disk1

Execute the following command to set the cluster information:
```
# mpaclcls -s 1
```
Refer to the Systemwalker Operation Manager Reference Guide for details on the mpaclcls command.
Execute the following command to reflect the security information in the files on the shared disk:
```
# mpcssave -s 1
```
Refer to the Systemwalker Operation Manager Reference Guide for details on the mpcssave command.

Unmount the shared disk.

# umount /disk1
# vgchange -a n /dev/vg01

For HACMP

In the examples that explain the settings procedure shown below, Subsystem 1 runs as a cluster system, the shared disk volume name is "datavg1", and the mount point to the shared disk is "/disk1".

Use the following procedure to configure the settings on the active node.

Mount the shared disk.
```
# varyonvg datavg1
# mount /disk1
```
Execute the following command to set the cluster information:
```
# mpaclcls -s 1
```
Refer to the Systemwalker Operation Manager Reference Guide for details on the mpaclcls command.
Execute the following command to reflect the security information in the files on the shared disk:
```
# mpcssave -s 1
```
Refer to the Systemwalker Operation Manager Reference Guide for details on the mpcssave command.
Unmount the shared disk.
```
# umount /disk1
# varyoffvg datavg1
```

Settings for dual node mutual standby configuration

The explanation provided here as an example is based on the cluster configuration shown in the following figure:

For PRIMECLUSTER (for Solaris or Linux version), Oracle Solaris Cluster

Mount the shared disk on Node A.
```
# mount /disk1 (*1)
```
*1
The command line for Linux version is as follows. (In this example, the device is "/dev/sdb1" the shared disk is "/disk1".)
```
# /bin/mount /dev/sdb1 /disk1
```
Set up cluster information by executing the following command on Node A.
```
# mpaclcls -s 1
```
Refer to the Systemwalker Operation Manager Reference Guide for details on the mpaclcls command.
Reflect security information in files on the shared disk by executing the following command on Node A.
```
# mpcssave -s 1
```
Refer to the Systemwalker Operation Manager Reference Guide for details on the mpcssave command.
Unmount the shared disk on Node A.
```
# umount /disk1 (*1)
```
*1
The command line for Linux version is as follows.
```
# /bin/umount /disk1
```
Mount the shared disk on Node B.
```
# mount /disk2 (*1)
```
*1
The command line for Linux version is as follows. (In this example, the device is "/dev/sdb2" the shared disk is "/disk2".)
```
# /bin/mount /dev/sdb2 /disk2
```
Set up cluster information by executing the following command on Node B.
```
# mpaclcls -s 2
```
Refer to the Systemwalker Operation Manager Reference Guide for details on the mpaclcls command.
Reflect security information in files on the shared disk by executing the following command on Node B.
```
# mpcssave -s 2
```
Refer to the Systemwalker Operation Manager Reference Guide for details on the mpcssave command.
Unmount the shared disk on Node B.
```
# umount /disk2 (*1)
```
*1
The command line for Linux version is as follows.
```
# /bin/umount /disk2
```

For MC/ServiceGuard

Mount the shared disk on Node A.

# vgchange -a e /dev/vg01
# mount /disk1

Set up cluster information by executing the following command on Node A.
```
# mpaclcls -s 1
```
Refer to the Systemwalker Operation Manager Reference Guide for details on the mpaclcls command.
Reflect security information in files on the shared disk by executing the following command on Node A.
```
# mpcssave -s 1
```
Refer to the Systemwalker Operation Manager Reference Guide for details on the mpcssave command.

Unmount the shared disk on Node A.

# umount /disk1
# vgchange -a n /dev/vg01

Mount the shared disk on Node B.

# vgchange -a e /dev/vg02
# mount /disk2

Set up cluster information by executing the following command on Node B.
```
# mpaclcls -s 2
```
Refer to the Systemwalker Operation Manager Reference Guide for details on the mpaclcls command.
Reflect security information in files on the shared disk by executing the following command on Node B.
```
# mpcssave -s 2
```
Refer to the Systemwalker Operation Manager Reference Guide for details on the mpcssave command.

Unmount the shared disk on Node B.

# umount /disk2
# vgchange -a n /dev/vg02

For HACMP

Mount the shared disk on Node A.
```
# varyonvg datavg1
# mount /disk1
```
Set up cluster information by executing the following command on Node A.
```
# mpaclcls -s 1
```
Refer to the Systemwalker Operation Manager Reference Guide for details on the mpaclcls command.
Reflect security information in files on the shared disk by executing the following command on Node A.
```
# mpcssave -s 1
```
Refer to the Systemwalker Operation Manager Reference Guide for details on the mpcssave command.
Unmount the shared disk on Node A.
```
# umount /disk1
# varyoffvg datavg1
```
Mount the shared disk on Node B.
```
# varyonvg datavg2
# mount /disk2
```
Set up cluster information by executing the following command on Node B.
```
# mpaclcls -s 2
```
Refer to the Systemwalker Operation Manager Reference Guide for details on the mpaclcls command.
Reflect security information in files on the shared disk by executing the following command on Node B.
```
# mpcssave -s 2
```
Refer to the Systemwalker Operation Manager Reference Guide for details on the mpcssave command.
Unmount the shared disk on Node B.
```
# umount /disk2
# varyoffvg datavg2
```

Settings for automatically reflecting calendar information and service/application startup information

The procedure for automatically reflecting calendar information is described below.

1) Enable the automatic reflection function

Information

In the operation of registering Systemwalker Operation Manager with a cluster system, the automatic reflection function allows you to automatically synchronize the setting information of the calendar and the service application startup function between active node and standby node by using Systemwalker Operation Manager daemons.

Enable the automatic reflection function for calendar information and service/application startup information.

Enable the automatic reflection function for calendar information and service/application startup information by executing the calsetcluster command on all the nodes (active and standby) on the cluster system.

The following examples show how this command is executed.

For PRIMECLUSTER (for Solaris or Linux version), Oracle Solaris Cluster
[Example where all information is automatically reflected with a 1:1 active/standby configuration]
```
# /opt/FJSVjmcal/bin/calsetcluster -type s
```
[Example where all information is automatically reflected with a N:1 active/standby configuration]
```
# /opt/FJSVjmcal/bin/calsetcluster -type n
```
[Example where all information is automatically reflected with a dual node mutual standby configuration]
```
# /opt/FJSVjmcal/bin/calsetcluster -type e
```
[Example where all information is automatically reflected with a cascading configuration]
```
# /opt/FJSVjmcal/bin/calsetcluster -type c
```
For MC/ServiceGuard
[Example where all information is automatically reflected with a 1:1 active/standby configuration]
```
# /opt/FHPjmcal/bin/calsetcluster -type s
```
[Example where all information is automatically reflected with a N:1 active/standby configuration]
```
# /opt/FHPjmcal/bin/calsetcluster -type n
```
[Example where all information is automatically reflected with a dual node mutual standby configuration]
```
# /opt/FHPjmcal/bin/calsetcluster -type e
```
For HACMP
[Example where all information is automatically reflected with a 1:1 active/standby configuration]
```
# /opt/FAIXjmcal/bin/calsetcluster -type s
```
[Example where all information is automatically reflected with a dual node mutual standby configuration]
```
# /opt/FAIXjmcal/bin/calsetcluster -type e
```

Refer to the Systemwalker Operation Manager Reference Guide for details on the calsetcluster command.

2) Set up the automatic reflection hosts

Define the hosts where calendar information and service/application startup information will be reflected in the calcphost.def definition file on the active and standby nodes.

When performing automatic reflection using an IP address (hereinafter referred to as "specific IP address") different from the local host IP address (acquired by name resolution, for example from the hosts file) on the nodes that use multiple physical IP addresses, perform steps 3 to 5 on those nodes that use a "specific IP address". The following steps 3 to 5 are not required if using the local host IP address (acquired by name resolution, for example from the hosts file).

Open the calcphost.def definition file with a text editor such as vi.
- For PRIMECLUSTER (for Solaris or Linux version), Oracle Solaris Cluster
```
/var/opt/FJSVjmcal/etc/calcphost.def
```
- For MC/ServiceGuard
```
/opt/FHPjmcal/etc/calcphost.def
```
- For HACMP
```
/opt/FAIXjmcal/etc/calcphost.def
```
Enter the physical IP addresses or the host names corresponding to the physical IP addresses for all the nodes on the cluster system. Make sure that the definition files include the host name or physical IP address of the local node.
Also, specify the same content for the calcphost.def definition file in all nodes that configure cluster systems.
When setting a "specific IP address" as the automatic reflection destination without using a "main IP address", specify the physical IP address to be used as the "specific IP address" instead of specifying a host name.
Open the myhostip.def definition file with a text editor such as vi in nodes that will use a "specific IP address".
- For PRIMECLUSTER (for Solaris or Linux version), Oracle Solaris Cluster
```
/var/opt/FJSVjmcal/etc/myhostip.def
```
- For MC/ServiceGuard
```
/opt/FHPjmcal/etc/myhostip.def
```
- For HACMP
```
/opt/FAIXjmcal/etc/myhostip.def
```

Enter the physical IP address that will be used as the "specific IP address".

XXX.XXX.XXX.XXX

Note: Only use alphanumeric characters, or "." and ":" as separators.
     Do not include characters other than those listed above such as leading or trailing blank spaces or newline codes.

Set access rights in the myhostip.def definition file.
Set write permissions for administrators and read permissions for all users.
```
# chmod 644 /var/opt/FJSVjmcal/etc/myhostip.def
```

To enable the contents of the definition file, start the calendar daemon.

Refer to the Systemwalker Operation Manager Reference Guide for details on the calcphost.def definition file.

Note

To apply automatic reflection to nodes that are already operating, first distribute policy to standardize the calendar information and service/application startup information on all the nodes. The information extracted by the policy is the information that will be subject to automatic reflection (either calendar information or service/application startup information or both).

Note

myhostip.def definition file

The physical IP address specified in the myhostip.def definition file must exist in the calendar reflection destination host calcphost.def definition file. Also if the IP address does not exist or the specification format is invalid, automatic reflection of calendar information will be invalid.
The myhostip.def definition file is not subject to backup and restore by Systemwalker Operation Manager. If necessary, manually perform backup or restore.