10.5.2 Supplement on Cluster Application and Resource Deletion

If the "userApplication Configuration Wizard" screen is used to create new script files for Cmdline resource creation, follow the procedures below to delete each script which is used for the Cmdline resource.

Note

Before deleting scripts, you need to delete the Cmdline resources.

Each script is stored in the following directories:

/etc/opt/FJSVwvucw/scripts/start
/etc/opt/FJSVwvucw/scripts/stop
/etc/opt/FJSVwvucw/scripts/check

Execute the deletion command as follows:

  # cd /etc/opt/FJSVwvucw/scripts/start

  # rm <script name to be deleted>

  # cd /etc/opt/FJSVwvucw/scripts/stop

  # rm <script name to be deleted>

  # cd /etc/opt/FJSVwvucw/scripts/check

  # rm <script name to be deleted>

Also, if you set up a dependency between RMS and the SMF service as described in "6.12.4 SMF Service Dependency Configurations," execute the procedures below on each node in the cluster and delete the set up dependency relationship.

Check the dependency name used during configuration.

Execute the command below to check the dependency name used during configuration.

# svccfg export svc:/milestone/smawrhvto | grep -w "dependency\ name" | grep -vw multi-user

Example

# svccfg export svc:/milestone/smawrhvto | grep -w "dependency\ name" | grep -vw multi-user
<dependency name='fuji' grouping='optional_all' restart_on='none' type='service'>

The value displayed in "dependency name=" on each line is the dependency name used during configuration.

Delete the dependency.

After RMS has been stopped, execute the following procedure and delete the dependency.

# /usr/sbin/svccfg -s svc:/milestone/smawrhvto delpg <dependency name>
# /usr/sbin/svcadm refresh svc:/milestone/smawrhvto

10.5.2.2 Supplement on Fsystem resource deletion

Restore the environment files related to the mount point to the original state.

Note

Before deleting the environment files, you need to delete the Fsystem resources.

Deleting mount point entries that were added at prerequisites

Delete the relevant entry from the following environment files that were edited as part of the "Prerequisites."

/etc/vfstab.pcl
/etc/dfs/dfstab.pcl

When an NFS Lock Failover was used

If the NFS Lock Failover functions are no longer used due to deletion of the Fsystem resources, take the following steps:

Procedure:

Change the settings for shared information with the configuration.
Change the value of NFSLockFailover to "no."

See

See "Editing global settings in Configuration."

10.5.2.3 Supplement on Gds resource deletion

Delete the disk class if it is no longer necessary.

See "PRIMECLUSTER Global Disk Services Configuration and Administration Guide."

10.5.2.4 Supplement related to Gls resource deletion

Delete the takeover network definition if it is no longer necessary.

See "5.4 Deleting configuration for Cluster System" in "PRIMECLUSTER Global Link Services Configuration and Administration Guide: Redundant Line Control Function."

10.5.2.5 Supplement on takeover network resource deletion

Restore the environment files related to the takeover network to the original state.

When node name takeover was used

Change the node name in the cluster system used for node name takeover. To bring the system state back to the state before the change, take the following steps on all the nodes where node name takeover resources were created.

Operation

Execute the command on all the nodes, as follows:
```
# setuname -n fuji2
```
fuji2 is used as the node name in this example.
Reboot the system.
```
# /usr/sbin/shutdown -g0 -i6 -y
```

10.5.2.6 Supplement on procedure resource deletion

Delete the procedure resource from the cluster resource management facility when it is no longer necessary.

See "E.3 Deleting a Procedure Resource."

10.5.2.7 Supplement on process monitoring resource deletion

Delete scripts that were created automatically if the stop command of the process monitoring resource was used.

Note

Before deleting the created scripts, you need to delete the process monitoring resources.

Each script is stored in the /etc/opt/FJSVwvucw/scripts/process_stop directory.

Execute the deletion command as shown below.

# cd /etc/opt/FJSVwvucw/scripts/process_stop
# rm ./*

10.5.2.8 Supplement on line switching unit resource deletion

Delete the line switching unit resource from the cluster resource management facility if it is not necessary.

See the manual pages of the "cldelswursc(1M)" command.

10.5.2.9 ISV resource deletion procedure (supplement)

For information on the deletion procedure, see the manuals provided with each product.

10.5.2.10 Supplement on I/O fencing resource deletion

When deleting the I/O fencing resource from the cluster application, you need to implement the following setting changes after deleting the resource.

When deleting the I/O fencing resource from all the cluster applications
- Change the setting to use XSCF SNMP when you use ICMP for the shutdown agent.
- Reboot the system after deleting the line of "SDX_VM_IO_FENCE=on" in the /etc/opt/FJSVsdx/sdx.cf file you have set in "3.2.4 Setting I/O Fencing Function of GDS."
When the I/O fencing resource is used continuously by other cluster applications
- Delete the Gds resource from the cluster application you have deleted the I/O fencing resource.

Note

The configuration after deleting the I/O fencing resource must match the support configuration in "2.2.1.1.1 Cluster System Configuration in Oracle VM Server for SPARC Environment."

See

For details of the I/O fencing function and ICMP shutdown agent, see "2.2.3 I/O Fencing Function."