Install required software as described in the section “Installing Required Software.”

Check the setting of two important NVRAM variables as described in the section “Setting NVRAM Variables.”

Create the XLV logical volumes and XFS filesystems required by the highly available applications you plan to run on the cluster. See the section “Creating XLV Logical Volumes and XFS Filesystems.”

Configure the network interfaces on both nodes using the procedure in the section “Configuring Interfaces.”

Configure the serial ports used on each node for the serial connection to the other node by following the procedure in the section “Configuring the Serial Port.”

If the IRIS FailSafe NFS option is used, perform necessary NFS configuration as described in the section “Configuring NFS Filesystems.”

If the IRIS FailSafe Web option is used, configure Netscape servers as described in the section “Configuring a Netscape Server.”

When you are ready configure the nodes so that IRIS FailSafe software starts up when they are rebooted, follow the instructions in the section “Configuring IRIS FailSafe On.”

On one node, follow normal software installation procedures and use inst to install required subsystems that aren't already installed. The required subsystems are listed in Table 3-1. (See “Related Documentation” for guides about inst).

On the same node, install required patches. See the release notes for IRIS FailSafe (called ha_base, ha_nfs, and ha_www) for lists of the required patches.

Install software and patches on the other node by repeating steps 1 and 2 on the other node.

If you are using plexed XLV logical volumes, install a disk plexing license on each node.

For IRIX 5.3, the license is a NetLS license and it is installed in a non-standard location, /etc/nodelock. For more information, see the Disk Plexing release notes (called plexing).

For IRIX 6.2, the license is a FLEXlm license and installed in /etc/flexlm/license.dat. For more information see the FLEXlm End User Manual.

If you are using plexed XLV logical volumes, verify that the license has been successfully installed by entering these commands on each node:

# xlv_mgr
xlv_mgr> show config
...
Plexing license: present
...
xlv_mgr> quit

The line shown about the plexing license indicates that the plexing license has been successfully installed.

If possible, add every IP address, IP name, and IP alias for both nodes to /etc/hosts on one node.

For example:

190.0.2.1 xfs-ha1.company.com xfs-ha1
190.0.2.3 stocks
190.0.3.1 priv-xfs-ha1
190.0.2.2 xfs-ha2.company.com xfs-ha2
190.0.2.4 bonds
190.0.3.2 priv-xfs-ha2

Add all of the IP addresses from step 1 to /etc/hosts on the other node in the cluster.

For IP addresses, IP names, and IP aliases that you did not add to /etc/hosts on the nodes, verify that they are in the NIS database by entering this command for each address:

# ypmatch address hosts
190.0.2.1 xfs-ha1.company.com xfs-ha1

address is an IP address, IP name, or IP alias. If ypmatch reports that address doesn't match, it must be added to the NIS database.

On one node, add that node's public and private interfaces and their fixed IP addresses to the file /etc/config/netif.options (high availability IP addresses are not added to the netif.options file).

For the example in Figure 3-1, the public interface name and IP address lines are:

if1name=ec0
if1addr=$HOSTNAME

$HOSTNAME is an alias for an IP address that appears in /etc/hosts.

If there are additional public interfaces, their interface names and IP addresses appear on lines like these:

if2name=
if2addr=

In the example, the private interface name and IP address are:

if3name=ec3
if3addr=priv-$HOSTNAME

The private interface IP address in this example, priv-$HOSTNAME, is an alias for an IP address that appears in /etc/hosts.

If there are more than eight interfaces on the node, change the value of if_num to the number of interfaces. For less than eight interfaces (as in the example in Figure 3-1) the line looks like this:

if_num=8

Repeat steps 4 and 5 on the other node.

Edit the file /etc/config/routed.options on each node so that the routes are not shown over the private network (routing is turned off). The content of /etc/config/routed.options should be:

-h -q

These options are required for IRIS FailSafe to function correctly.

Verify that IRIS FailSafe is chkconfig'd off on each node:

# chkconfig 
        Flag                 State
        ====                 =====
        ...
        failsafe             off
        ...

If failsafe is on on either node, enter this command on that node:

# chkconfig failsafe off 

Reboot both nodes to put the network configuration into effect.

To configure your e-mail interface so that you can receive notification of cluster transitions, set up an e-mail alias on each node that includes at least one user outside the IRIS FailSafe cluster and an alias on the other node that can direct notification to an outside user. For example, in /usr/lib/aliases on each node, add

fsafe_admin:outside,outside@xfs-ha1.company.com 
outside:operations@console.company.com

If the nodes use NIS (yp is chkconfig'ed on) or the BIND domain name server (DNS), switching to local name resolution is recommended. Create or modify the file /etc/resolv.conf so that local is listed first for the hostresorder keyword (the order of nis and bind is up to you):

hostresorder local nis bind 

If FDDI is being used, finish configuring the new FDDI station, as explained in Chapter 2 of the FDDIXpress Administration Guide. If you are familiar with the procedures, follow the section “Quick and Easy Configuration Instructions” in that chapter. Verify the FDDI connection as explained in Chapter 2 of the FDDIXpress Administration Guide.

Determine which port is used for the reset serial line (see the section “Serial Port Configuration” in Chapter 2).

Open the file /etc/inittab for editing.

Find the line for the port by looking at the comments on the right for the port number from step 1.

Change the third field of this line to off, for example:

t2:23:off:/sbin/getty -N ttyd2 co_9600          # port 2

Save the file.

Enter these commands to make the change take effect:

# killall getty
# init q

Create or identify the filesystems to be failed over. They must follow the guidelines in the section “Planning Filesystems” in Chapter 2.

On each node that exports filesystems, create the statmon directory on any one of the exported filesystems on a shared disk. The name of the directory must be statmon.

For nodes that are running IRIX 5.3, skip the remaining steps in this procedure; they apply only to nodes that are running IRIX 6.2.

On one node, open the file /etc/config/statd.options for editing (it may be new).

Put -h on the first line of the file:

-h 

Close the file.

Set the owner, group, and mode of the file, if you just created it:

# chown root.sys /etc/config/statd.options
# chmod 644 /etc/config/statd.options

Repeat steps 4 through 7 on the other node.

Because Netscape server installation requires that the interface be accessible from a Netscape browser, ifconfig the interface to the public network up by entering this command:

# /usr/etc/ifconfig interface alias ip-address netmask netmask

interface is the interface to be used access the server, ip-address is a high availability IP address for the interface, and netmask is the netmask of the IP address.

To install the first Web server (stocks), enter

# cd /usr/netscape/httpd/install
# ./ns-setup

Start the Netscape browser on a workstation and open the Netscape server's configuration page (see the output of the ns-setup command for the URL).

Click the Server Config button and enter information according to the prompts. An example of the server information is as follows:

Click the Document Config button and enter information according to the prompts. An example of the information is as follows:

Click the Admin Config button and enter information according to the prompts. Exit the page.

Copy the contents of the server's root (/usr/ns-home/httpd-80.190.2.3) and document root (/usr/ns-home/stocks_root) to a filesystem on a shared disk.

Replace the original directory with symbolic links to the directories on the shared disk.