This section describes the steps you take to customize HACWS and verify its function.
Step 29: Set up the HACWS configuration requires each control workstation be addressable by its host name. If the name of the primary control workstation is dutchess.xyz.com, the backup control workstation must be able to communicate with the primary control workstation using the name dutchess.xyz.com. If the name of the backup control workstation is ulster.xyz.com, the primary control workstation must be able to communicate with the backup control workstation using the name ulster.xyz.com.
If either control workstation is not addressable by its host name, you need to enter the appropriate ifconfig command to cause the required name to be temporarily configured as a network interface.
An example of an ifconfig command is as follows:
ifconfig en0 dutchess.xyz.com netmask 255.255.255.192 up
The ifconfig options you need to use are specific to your site. Refer to the ifconfig man page for more information.
Configure the primary and backup control workstation as an HACWS configuration.
| If using: | Do this:. |
|---|---|
| SMIT |
|
| install_hacws | Enter:
/usr/sbin/hacws/install_hacws -p \ primary_hostname -b backup_hostname -s |
Identify the HACMP pre- and post-event scripts provided by HACWS to
HACMP.
| If using: | Do this:. |
|---|---|
| SMIT |
|
| spcw_addevents | Enter:
/usr/sbin/hacws/spcw_addevents |
If you want to further customize cluster event processing, follow the instructions in the "Managing a high availability control workstation" chapter in PSSP: Administration Guide.
In Step 2: Plan network configuration you planned the network configuration for both control workstations. If your configuration requires IP address aliases, you need to provide the appropriate commands to HACWS. If these commands already existed somewhere else before you installed HACWS, remove them from the old location.
If you need to use an IP address alias to configure the host name of the backup control workstation as a network interface, edit the /etc/rc.backup_cw_alias script on the backup control workstation. This script only runs on the backup control workstation. See the comments in the script for more details. If you make any changes to the /etc/rc.backup_cw_alias script, you must also edit the /etc/rc.net script so /etc/rc.backup_cw_alias runs when the backup control workstation boots. Use an editor to add the following line near the end of the /etc/rc.net file:
/etc/rc.backup_cw_alias
If you need to use an IP address alias to configure a network interface for an SP system partition (SP Switch only), or if you need to configure an IP address alias on the active control workstation for some other reason, edit the /spdata/sys1/hacws/rc.syspar_aliases script. This script runs on the active control workstation before it starts the control workstation services. See the comments in the script for more details.
Verify the HACWS configuration.
| If using: | Do this:. |
|---|---|
| SMIT |
|
| hacws_verify | Enter:
/usr/sbin/hacws/hacws_verify |
Verify the connections between the backup control workstation and the
frames. Run this step from the backup control workstation.
| If using: | Do this:. |
|---|---|
| SMIT |
|
| spcw_verify_cabling | Type:
/usr/sbin/hacws/spcw_verify_cabling |
Reboot the primary control workstation. After it finishes booting, reboot the backup control workstation.
Follow the instructions in the "Starting and stopping cluster services" chapter of the HACMP for AIX: Administration Guide to start cluster services on the primary control workstation.
Using SMIT:
To make sure you installed the HACWS correctly, verify that the control workstation services exist and can move between the primary and backup control workstations.
Make sure the control workstation services come up on the primary control workstation. Use the following procedure to test them:
df /spdata
/usr/lpp/ssp/bin/SDRGetObjects SP
The test is successful if you receive output. You can tell whether the startup of the control workstation services has completed by issuing the command:
grep "SPCW_APPS COMPLETE" /tmp/hacmp.out
Once control workstation services have completely started on the primary control workstation, start cluster services on the backup control workstation by following the procedure described in Step 35: Start cluster services on the primary control workstation.
After cluster services have started on the primary control workstation, direct HACMP to move control workstation services to the backup control workstation.
Using SMIT:
This should stop cluster services on the primary control workstation and move control workstation services to the backup control workstation. Make sure control workstation services move to the backup. On the backup control workstation, follow the procedure described in Verify the control workstation services.
When you are finished testing, restart cluster services on the primary control workstation by following Step 35: Start cluster services on the primary control workstation.
grep "SPCW_APPS COMPLETE" /tmp/hacmp.out
Once the startup of the control workstation services has completed on the backup control workstation, you can restart cluster services on the primary control workstation and then move control workstation services back to the primary control workstation. When you have completed testing, make sure to restart cluster services on both control workstations.