7.5.1 Notes common to each type of cluster software
This subsection provides notes common to each type of cluster software.
- Organization of this subsection
(1) About starting and stopping SSO
If SSO is in a cluster system, use cluster software to start or stop SSO.
However, if abnormality of SSO (failure of an SSO process) is not set as a failover trigger, you can use the ssostop and ssostart commands to restart only SSO. To restart only SSO when abnormality of SSO is set as a failover trigger, see (2) About restarting the SSO processes.
(2) About restarting the SSO processes
You might need to restart the SSO processes after changing action definition files or user resource configuration files. However, stopping SSO processes that are monitored by using the cluster control script of SSO causes a failover to occur.
To stop SSO processes that are monitored by using the cluster control script without causing a failover to occur, create the monitoring interrupt file to enter maintenance mode, and then use the ssostop command to stop SSO.
After that, to start the SSO processes, use the ssostart command to start SSO, and then delete the monitoring interrupt file to exit maintenance mode.
You can create the monitoring interrupt file with a size of 0 bytes.
The following table lists the location of the monitoring interrupt file.
|
Name of the monitoring interrupt file |
Directory |
|---|---|
|
ha_maintenance |
$SSO_TMP |
Note that you need to perform the above procedure only when SSO processes are monitored by using the cluster control script of SSO.
(3) About maintenance mode of the NNMi resource
Before you place the NNMi resource in maintenance mode, check whether SSO satisfies the conditions shown below. If these conditions are satisfied, before stopping NNMi, create the monitoring interrupt file to place the SSO resource in maintenance mode.
-
In the cluster control script, processes other than the ssospmd command have been added as monitoring targets of SSO.
-
In the ssospmd action definition file, the nnm-coop-policy: key is set to 0.
The SSO resource depends on the NNMi resource. Therefore, when you enable maintenance mode, place the SSO resource in maintenance mode, and then place the NNMi resource in maintenance mode. When you disable maintenance mode, do so for the NNMi resource, and then for the SSO resource.
(4) About modifying the cluster control script
Before you modify the SSO cluster control script, deactivate the SSO resource.
If you want to modify the cluster control script without deactivating the SSO resource, modify the cluster control script on the host on which no SSO processes are running, and then do so again on the other host after switching the active node to the other node.
You cannot modify the cluster control script on the host on which the SSO resource is activated.
(5) About changing the logical IP address
(a) In the basic configuration
If you change the logical IP address of SSO, do so before changing the virtual IP address of NNMi (before deactivating the NNMi resource group).
To change the logical IP address of SSO, specify the following arguments in the cluster environment setting command ssoclustersetup.vbs or ssoclustersetup:
|
Argument |
Value to be specified |
|---|---|
|
First argument |
-defset (specifies that the action definition file is to be set) |
|
Second argument |
New logical IP address (IPv4 address or IPv6 address) |
Note that you cannot change the logical IP address when the NNMi resource group is inactive.
The following shows an example of the command that changes the IPv4 logical IP address in Windows:
cscript.exe $SSO_BIN\ssoclustersetup.vbs -defset new-logical-IP-address
The following shows an example of the command that changes the IPv6 logical IP address in Linux:
ssoclustersetup -defset new-logical-IP-address
After executing the above command, perform the following procedure:
-
Create the monitoring interrupt file (only for WSFC and HA Monitor).
For details about how to create the file, see (2) About restarting the SSO processes.
-
Perform the procedure for changing the IP address of the local server on which SSO is running.
For details, see 2.9.1(7) Changing the host name or IP address of the local host on which SSO runs. Note the following points:
-
In the procedure described in subsection 2.9.1(7), replace "IP address" with "logical IP address".
-
Perform step 8 before step 6.
-
In step 6, deactivate the resource group. In step 9, activate the resource group (for Veritas Cluster Server only).
The following command deactivates the resource group:
# hagrp -offline resource-group-name -sys name-of-the-host-to-be-stopped
The following command activates the resource group:
# hagrp -online resource-group-name -sys name-of-the-host-to-be-activated
-
-
Delete the monitoring interrupt file (only for WSFC and HA Monitor).
If you changed the virtual IP address of NNMi before changing the logical IP address of SSO, perform the following procedure:
-
Create the monitoring interrupt file (only for WSFC and HA Monitor).
For details about how to create the file, see (2) About restarting the SSO processes.
-
In the Process Configuration window, delete all entries of the local host from List of monitoring servers.
-
In the Resource Configuration window, delete all collection conditions for the entries of the local host.
-
Stop all resource collection and process monitoring tasks.
-
Close all windows for resource collection and process monitoring.
-
Delete all resource collection databases of SSO.
If necessary, before you delete the databases, use the ssoextractlog command to extract and save the data that is in the databases.
-
Execute the cluster environment setting command (ssoclustersetup.vbs or ssoclustersetup) with the -defset argument specified to change the logical IP address.
-
If a host whose logical IP address is changed is defined in the following definition files, change the relevant definitions in the files:
SSO. definition files
-
Event destination definition file (ssodest.conf)
-
Monitoring server definition file
APM. definition files
-
Startup event destination definition file (apmdest.conf)
-
Proxy definition file (apmproxy.conf)
A host whose logical IP address is changed might be defined as a destination of SSO events (in the ssodest.conf file) and a destination of SNMP agent events. Check whether the host is defined as those destinations on the local and remote hosts, and if defined, change the relevant definitions.
Note that in Linux, SNMP agent events are sent to the destinations set for trap-dest in /etc/SnmpAgent.d/snmpd.conf. In Windows, SNMP agent events are sent to the destinations displayed in Trap destinations in the Traps tab of the properties window for the service named SNMP Service.
-
-
Stop SSO.
In WSFC and HA Monitor:
Execute the ssostop command to stop SSO.
In Veritas Cluster Server:
Execute the following command to deactivate the resource group:
# hagrp -offline resource-group-name -sys name-of-the-host-to-be-stopped
-
Start SSO.
In WSFC and HA Monitor:
Execute the ssostart command to start SSO with the new definitions applied.
In Veritas Cluster Server:
Execute the following command to activate the resource group:
# hagrp -offline resource-group-name -sys name-of-the-host-to-be-activated
-
Delete the monitoring interrupt file (only for WSFC and HA Monitor).
-
Respecify the conditions for resource collection and processing monitoring.
-
Start resource collection and process monitoring.
-
Re-create the report definition file.
(b) In the distributed configuration
To change the logical IP address of SSO, specify the following arguments in the cluster environment setting command ssoclustersetup.vbs or ssoclustersetup:
|
Argument |
Value to be specified |
|---|---|
|
First argument |
-defset (specifies that the action definition file is to be set) |
|
Second argument |
New logical IP address (IPv4 address or IPv6 address) |
The following shows an example of the command that changes the IPv4 logical IP address in Windows:
cscript.exe $SSO_BIN\ssoclustersetup.vbs -defset new-logical-IP-address
The following shows an example of the command that changes the IPv6 logical IP address in Linux:
ssoclustersetup -defset new-logical-IP-address
After executing the above command, perform the following procedure:
-
Create the monitoring interrupt file (only for WSFC and HA Monitor).
For details about how to create the file, see (2) About restarting the SSO processes.
-
Perform the procedure for changing the IP address of the local server on which SSO is running.
For details, see 2.9.1(7) Changing the host name or IP address of the local host on which SSO runs. Note the following points:
-
In the procedure described in subsection 2.9.1(7), replace "IP address" with "logical IP address". When changing IP addresses, change the logical IP addresses of the resources in the SSO resource group by using the procedure that is appropriate for the cluster software.
-
Perform step 8 before step 6.
-
In step 6, deactivate the resource group. In step 9, activate the resource group (for Veritas Cluster Server only).
The following command deactivates the resource group:
# hagrp -offline resource-group-name -sys name-of-the-host-to-be-stopped
The following command activates the resource group:
# hagrp -online resource-group-name -sys name-of-the-host-to-be-activated
-
-
Delete the monitoring interrupt file (only for WSFC and HA Monitor).
(6) About renaming the logical host
(a) In the basic configuration
Before you rename the logical host, you must delete and re-create the NNMi resource group.
Release the SSO cluster environment, rename the NNMi virtual host, and then reconfigure the SSO cluster environment.
To rename the SSO logical host:
-
Release the SSO cluster system environment.
For details about releasing a cluster system environment, see 7.4 Releasing the SSO cluster system environment. Also note the following points:
-
Do not delete the SSO cluster control script because it is reused after the environment is reconfigured.
-
To inherit the settings, use the cluster environment setting command to undo setup on only the active node. You do not need to undo cluster system setup on the standby node.
-
-
Remove the HA (high-availability) configuration of NNMi.
For details, see the Release Notes of NNMi.
-
Rename the logical host.
For details, see the Release Notes of NNMi.
-
Set up an HA (high-availability) configuration of NNMi.
For details, see the Release Notes of NNMi.
-
Reconfigure the SSO cluster system environment.
For details about reconfiguring a cluster system environment, see 7.2 Configuring an SSO cluster system environment. Also note the following points:
-
Reuse the SSO cluster control script that was used before the cluster system environment was released.
-
You do not need to execute the ssonnmsetup command.
-
You do not need to set up a cluster system on the standby node.
-
-
Create the monitoring interrupt file (only for WSFC and HA Monitor).
For details about how to create the file, see (2) About restarting the SSO processes.
-
Perform the procedure for changing the host name of the local server on which SSO is running.
For details, see 2.9.1(7) Changing the host name or IP address of the local host on which SSO runs. Note the following points:
-
In the procedure described in subsection 2.9.1(7), replace "horst name" with "logical host name".
-
Perform step 8 before step 6.
-
In step 6, deactivate the resource group. In step 9, activate the resource group (for Veritas Cluster Server only).
The following command deactivates the resource group:
# hagrp -offline resource-group-name -sys name-of-the-host-to-be-stopped
The following command activates the resource group:
# hagrp -online resource-group-name -sys name-of-the-host-to-be-activated
-
-
Delete the monitoring interrupt file (only for WSFC and HA Monitor).
(b) In the distributed configuration
To rename the SSO logical host:
-
Create the monitoring interrupt file (only for WSFC and HA Monitor).
For details about how to create the file, see (2) About restarting the SSO processes.
-
Perform the procedure for changing the host name of the local server on which SSO is running.
For details, see 2.9.1(7) Changing the host name or IP address of the local host on which SSO runs. Note the following points:
-
In the procedure described in subsection 2.9.1(7), replace "host name" with "logical host name". When changing host name, change the logical host name of the resources in the SSO resource group by using the procedure that is appropriate for the cluster software.
-
Perform step 8 before step 6.
-
In step 6, deactivate the resource group. In step 9, activate the resource group (for Veritas Cluster Server only).
The following command deactivates the resource group:
# hagrp -offline resource-group-name -sys name-of-the-host-to-be-stopped
The following command activates the resource group:
# hagrp -online resource-group-name -sys name-of-the-host-to-be-activated
-
-
Delete the monitoring interrupt file (only for WSFC and HA Monitor).
(7) About overwrite installation or upgrade installation of NNMi (in the basic configuration)
Before you install NNMi as an overwrite installation or upgrade installation, deactivate the SSO resource, and then deactivate the NNMi resource.
Note that the SSO resource is activated during the procedure for overwrite-installing NNMi.
(8) About backup and restoration
When you perform backup or restoration, execute the ssobackup or ssorestore command on the active node only. Do not execute the command on the standby node. If you do so, the command results in a run-time error with return code -1.
Note that before you execute the ssorestore command, enter maintenance mode and stop SSO.
(9) About resource group members
We recommend that the NNMi resource group in the basic configuration or the SSO resource group in a distributed configuration does not include other JP1 resources. The only exceptional JP1 resource is JP1/Base in cases where JP1 authentication is used.