Review the sections “Disk Configuration,” “Logical Volume Configuration,” and “Filesystem Configuration” in Chapter 2 of the IRIS FailSafe Administrator's Guide in preparation for the planning described in the remainder of this subsection.

Prepare a table similar to Table 1-1 that you will use to make notes about logical volumes and filesystems. It will have one column for each logical volume that is to be failed over. A minimum of two columns, one for the INFORMIX database and one for video storage is required. The table shown in Table 1-1 describes an example configuration that is used throughout this guide: there is one logical volume for the INFORMIX database and one for video storage.

Choose a shared disk that you will use for the logical volume for the WebFORCE MediaBase INFORMIX database. It must be at least 200 MB. In the table, enter the full pathname of the logical volume, a volume name (label) you will use in the IRIS FailSafe configuration file, the full pathname of the mount point you will use for the filesystem on that logical volume, and a filesystem name (label) you will use in the IRIS FailSafe configuration file.

Identify one or more shared disks for the logical volumes you will use to store WebFORCE MediaBase video images. See the section “Creating Mediabase Filesystems” in Chapter 4 of the WebFORCE MediaBase Administrator's Guide for more information. For each of the logical volumes, enter information into your table.

Review the sections “Disk Configuration,” “Logical Volume Configuration,” and “Filesystem Configuration” in Chapter 2 of the IRIS FailSafe Administrator's Guide again and perform any additional required planning.

Review the planning described in the section “Network Interface and IP Address Configuration” in Chapter 2 of the IRIS FailSafe Administrator's Guide in preparation for the planning described in this subsection.

Prepare a table similar to Table 1-2 that you will use to make notes about IP names that are important to IRIS FailSafe and WebFORCE MediaBase. A completed table for an example which is used throughout this guide is shown in Table 1-2.

In the IP Name column of the table, enter the hostnames of the nodes in the IRIS FailSafe cluster. The hostname of the node that is to be the primary server should be entered on the first line and the hostname of the backup server should be entered on the second line.

Choose an IP name for the WebFORCE MediaBase server. This is the name that WebFORCE MediaBase clients will use to access the WebFORCE MediaBase server. It will be a high-availability IP address. As an example, the name mbasesrvr is used in the remainder of this procedure. The fully qualified version of the name is mbasesrvr.company.com.

Enter the fully qualified form of each IP name and the IP address for each IP name into the table.

Perform all additional planning described in the section “Network Interface and IP Address Configuration” in Chapter 2 of the IRIS FailSafe Administrator's Guide.

Identify which Netscape server, FastTrack^™ or Enterprise^™, will be used and how it will be installed. Enter these commands to determine if the FastTrack or Enterprise server is already installed:

# versions ns_fasttrack
# versions ns_enterprise

If neither server is installed, you will need to install one of the servers (in step 4 of the section “Part 1. Check Prerequisites”).

Prepare a table similar to Table 1-3 that you will use for future reference. Table 1-3 contains two columns, one for Netscape FastTrack server names and one for Netscape Enterprise Server names. Your table will have just one column for the server you will use. Copy the entries shown in Table 1-3.

Verify that the IRIS FailSafe cluster hardware (two nodes and shared disk storage that have been cabled together) is available.

Verify that the shared disks are accessible to the node that is to be the primary WebFORCE MediaBase server by examining the cluster and looking at the output of the hinv command.

Verify that the basic IRIX subsystems that are required by WebFORCE MediaBase, eoe.sw.ipgate, eoe.sw.xfs, eoe.sw.xfsrt, eoe.sw.xlv, and fw_LWperl5.sw.perl, are installed on both nodes by using the versions command.

Verify that the software for a Netscape server is installed on both nodes. It can be either the FastTrack server, ns_fasttrack, or the Enterprise server, ns_enterprise. In addition, ns_admin must be installed. If you need to install a Netscape server, the FastTrack server software may be included on your “WebFORCE MediaBase Server 2.0” CD. Be sure to install all of the ns_fasttrack.sw.* subsystems.

Verify that IRIS FailSafe base software, ha.sw.base, and any required IRIS FailSafe patches are installed on both nodes.

Verify that you have access to required manuals:

Verify that you have installed all required FLEXlm licenses on both nodes. Depending upon your configuration, these may include licenses for:

On the primary node for WebFORCE MediaBase (mb1 in the example configuration) create an XLV logical volume on a shared disk. On the logical volume create an XFS filesystem for the WebFORCE MediaBase INFORMIX database and mount the filesystem. See the IRIX Admin: Disks and Filesystems guide for details.

For the example shown in Table 1-1 in the section “Planning Logical Volumes and Filesystems for WebFORCE MediaBase and Its INFORMIX DataBase” in this chapter, the logical volume would be /dev/xlv/inf_data and the filesystem would be mounted at /mbaseinf_data.

On the primary node, create the mount point for the filesystem from step 1 in this subsection, for example /mbaseinf_data, and mount the filesystem. Do not make an entry in /etc/fstab for this filesystem.

On the backup node, create the same mount point for the filesystem for the INFORMIX database, for example /mbaseinf_data.

On the primary node, create one or more XLV logical volumes on shared disks to store videos. Create XFS filesystems on these logical volumes. These logical volumes and filesystems have special requirements that are described in the WebFORCE MediaBase Administrator's Guide in the section “Creating Mediabase Filesystems” in Chapter 4 and in Appendix B, “Building XLV Logical Disk Volumes for MediaBase.”

For the example configuration used in this guide, the logical volume for videos is /dev/xlv/video and the filesystem mount point is /mbasevideo.

On the primary node, create a mount point for each filesystem from step 4 in this subsection, for example /mbasevideo, and mount the filesystems. Do not make an entry in /etc/fstab for these filesystems.

On the backup node, create the same mount points for the filesystems for the videos, for example /mbasevideo.

On the node on which the shared disk filesystems are mounted, insert the CD labeled “WebFORCE MediaBase Server 2.0” into the server's drive. Install the server software by entering these commands:

# inst -f /CDROM/dist
Inst> go
Inst> quit

Check that the date and time are set correctly.

Create a file named /usr/ocs/etc/hostname (this is the literal filename; do not substitute a hostname) and add this line of text:

mbasesrvr.company.com

Use the fully qualified IP name used by WebFORCE MediaBase clients (from step 4 in the section “Planning IP Names Used for WebFORCE MediaBase Nodes” in this chapter).

Edit the file /etc/hosts and add an entry for the IP name used by WebFORCE MediaBase clients. For example:

190.0.2.4 mbasesrvr

Edit the file /usr/mbase/informix/etc/onconfig and change this line:

ROOTPATH        /usr/mbase/informix/data/root_chunk

to

ROOTPATH       /mbaseinf_data/root_chunk

The new directory is the mount point for the INFORMIX database from step 3 in the section “Planning Logical Volumes and Filesystems for WebFORCE MediaBase and Its INFORMIX DataBase” in this chapter.

Make the IP name used by WebFORCE MediaBase clients active by entering this ifconfig command:

# /usr/etc/ifconfig interface alias mbasesrvr netmask netmask

interface is the interface to be used by clients to access the WebFORCE MediaBase server, mbasesrvr should be replaced by the IP name (IP alias) used by WebFORCE MediaBase clients, and netmask is the netmask of the IP name. interface and netmask were identified during step 6 of the section “Planning IP Names Used for WebFORCE MediaBase Nodes” in this chapter.

On the node on which the shared disk filesystems are mounted, run the configuration and initialization utility mbaseinit:

# /usr/mbase/sbin/mbaseinit

mbaseinit asks you to confirm some changes. Enter y to confirm changes. If instructed to do so, do reboot the node.

On the same node, make the IP name used by WebFORCE MediaBase clients active by enter the ifconfig command from step 6 in the section “Part 3. Install and Begin to Configure the WebFORCE MediaBase Server on One Node” again:

# /usr/etc/ifconfig interface alias mbasesrvr netmask netmask

On the same node, mount all of the filesystems on shared disks using the mount command. For example

# mount /dev/xlv/inf_data /mbaseinf_data
# mount /dev/xlv/video /mbasevideo

On the same node, run mbaseinit again:

# /usr/mbase/sbin/mbaseinit

If you are using any network interfaces in addition to IP, follow the instructions in the section “Installing and Configuring Network Interfaces” in Chapter 4 of the WebFORCE MediaBase Administrator's Guide on the node on which the shared disk filesystems are mounted.

Follow the instructions in the section “Configuring and Starting MediaBase” in Chapter 4 of the WebFORCE MediaBase Administrator's Guide to configure WebFORCE MediaBase. However, do not start WebFORCE MediaBase (step 4 of the procedure).

Verify that the installation of the WebFORCE Media Server is complete by performing the steps in the section “Verifying the Installation of the Server” in Chapter 4 of the WebFORCE MediaBase Administrator's Guide.

Make sure that the WebFORCE MediaBase and Guaranteed-Rate I/O (GRIO) licenses are valid by entering this command:

# /usr/mbase/sbin/mblicense 

Stop the INFORMIX database and WebFORCE MediaBase by entering these commands on the node that has access to the shared disks:

# /etc/init.d/informix stop
# /etc/init.d/mediabase stop

Move the filesystems on shared disks so that they are mounted on the other node (the backup node, mb2) by following the procedure in the section “Switching Shared Disks to the Other Node” in this chapter.

On the node that now has access to the shared disks, make the IP name used by WebFORCE MediaBase clients active by enter the ifconfig command from step 6 in the section “Part 3. Install and Begin to Configure the WebFORCE MediaBase Server on One Node” again:

# /usr/etc/ifconfig interface alias mbasesrvr netmask netmask

On the same node, follow the steps in these sections:

To help you determine some of the IRIS FailSafe timing parameters, perform a timing experiment on the node that has access to the filesystems on shared disks by giving the commands below. While executing the commands, time the two start commands (from the time at which you give the commands to the time at which the prompt following the command appears).

# /etc/init.d/mediabase stop
# /etc/init.d/informix stop
# /etc/init.d/informix start
# /etc/init.d/mediabase start
# /etc/init.d/informix stop
# /etc/init.d/mediabase stop

Add the two times, which are the startup times for WebFORCE MediaBase and INFORMIX, and the approximate length of time it takes to mount the shared filesystems (10 or 20 seconds is typical). This number is a starting point for values of the long-timeout and start-monitor-time parameters.

Perform the node configuration tasks described in Chapter 3, “Configuring Nodes for IRIS FailSafe,” of the IRIS FailSafe Administrator's Guide. Be sure to include the IP name you chose in the section “Planning IP Names Used for WebFORCE MediaBase Nodes” in this chapter as you perform steps 1 through 4 of the section “Configuring Network Interfaces” in Chapter 3.

Create a simple IRIS FailSafe configuration file by following the procedure in the section “Creating a Configuration File” in Chapter 4 of the IRIS FailSafe Administrator's Guide. Use only the ha.conf.system, ha.conf.interfaces, ha.conf.volumes, and ha.conf.filesystems template files. Include the IP names, logical volumes, and filesystems from the sections “Planning IP Names Used for WebFORCE MediaBase Nodes” and “Planning Logical Volumes and Filesystems for WebFORCE MediaBase and Its INFORMIX DataBase” in this chapter. Use the starting point value for long-timeout and start-monitor-time from step 1 in this subsection.

A complete example of /var/ha/ha.conf appears in Appendix A, “Example Configuration File.”

Test the basic IRIS FailSafe configuration by following the first five procedures in Chapter 5 of the IRIS FailSafe Administrator's Guide: “Testing the Serial Connection,” “Testing the Private Network,” “Testing the Public Network Interfaces,” “Testing Logical Volumes,” and “Testing Filesystems.”

Test the basic IRIS FailSafe operation by performing most of the procedures in the section “Testing System Behavior With IRIS FailSafe Running” in Chapter 5 of the IRIS FailSafe Administrator's Guide. Omit these two portions of the procedures:

Checking failover and starting IRIS FailSafe will be done in “Part 9. Test WebFORCE MediaBase With IRIS FailSafe.”

On the node that has access to the shared disks, append the configuration file template /var/ha/templates/ha.conf.mediabase to the end of the file /var/ha/ha.conf.

On the same node, get a list of the WebFORCE MediaBase services by entering this command:

# /usr/mbase/sbin/cksvcs

Using Example 1-1 and the information in Chapter 2, “Configuration File Blocks for WebFORCE MediaBase,” as a guide, enter information into the template at the end of /var/ha/ha.conf as follows:

---

Note: The ha.conf configuration file for a cluster supplying WebFORCE MediaBase as a high-availability service doesn't require any webserver or informix blocks.

---

application-class mediabase
{
        server-node = primary_node_label
}

mediabase any_label
{
        server-node = primary_node_label
        backup-node = backup_node_label
        check-services = ( "service1" "service2" ... )
        httpd-dir = /usr/ns-home/httpd-mediabase
        httpd-options-file = ns_fasttrack.options
        httpd-script = /etc/init.d/ns_fasttrack
}

action mediabase
{
        local-monitor = /var/ha/actions/ha_mbase_lmon
}

action-timer mediabase
{
        start-monitor-time = 120
        lmon-probe-time = 120
        lmon-timeout = 90
        retry-count = 2
}

A complete example of /var/ha/ha.conf appears in Appendix A, “Example Configuration File.”

Verify the format and contents of your configuration file by running the ha_cfgverify command:

# /usr/etc/ha_cfgverify

The messages output by this command are described in Appendix A, “Messages About Configuration File Errors,” in the IRIS FailSafe Administrator's Guide.

For warnings reported by ha_cfgverify, check manually that there are no errors in the configuration file.

Resolve errors reported by ha_cfgverify and rerun the command until it reports that the file has passed.

Copy the configuration file to /var/ha/ha.conf on the other node in the cluster and repeat steps 4 through 6 in this subsection.

If necessary, copy the final configuration file to other node to ensure that the copies of the configuration file on each node are identical.

On whichever node it is running, shut down WebFORCE MediaBase by entering this command:

# /etc/init.d/mediabase stop 

Prevent WebFORCE MediaBase from coming back up automatically by entering these commands on both nodes:

# chkconfig informix off
# chkconfig mediabase off

Start IRIS FailSafe on each node by entering this command on each node:

# /etc/init.d/failsafe start 

Test the operation of WebFORCE MediaBase from the client's perspective by following the procedure in step 5 of the section “Verifying the Installation of the Server” in Chapter 4 of the WebFORCE MediaBase Administrator's Guide. However, do not use the hostname of the WebFORCE Media Server; rather, use the IP name that clients are to use from the section “Planning IP Names Used for WebFORCE MediaBase Nodes” in this chapter.

Test IRIS FailSafe system operation by following the instructions in the section “Testing System Behavior With IRIS FailSafe Running” in Chapter 5 of the IRIS FailSafe Administrator's Guide. You can omit the steps in the subsection “Checking Normal Operation.” Each time you cause a fail over, verify that the backup node has become the WebFORCE MediaBase server by playing a video on a client, as in step 4 in this subsection. Remember to use the same IP name as in step 4.

---

Note: At the end of the testing procedure, IRIS FailSafe and WebFORCE MediaBase, including the Netscape server and INFORMIX database, are running and the cluster is configured so that everything starts up automatically each time the nodes boot.

---

On node A, unmount the filesystems on the shared disks by entering the umount command for each of the filesystems:

# umount mountpoint

mountpoint is a filesystem mount point.

On node A, shut down each of the logical volumes on the shared disks by entering xlv_shutdown commands:

# xlv_shutdown -n volname

volname is the last component of the device filename /dev/xlv/volname.

On node B, change the host of the logical volumes by entering the xlv_mgr command and entering a change command that lists each logical volume:

# xlv_mgr
> change nodename hostnameB volname1 volname2 ...
> quit 

hostnameB is the hostname of node B (use the simple hostname, not the fully qualified hostname) and volname1 and volname2 are the last components of the device filenames /dev/xlv/volname1 and /dev/xlv/volname2.

On node B, assemble the logical volumes by entering this command:

# xlv_assemble -l  

On node B, mount the filesystems on the shared disks by entering the mount command for each of the filesystems:

# mount device mountpoint

device is the full pathname of the device filename and mountpoint is a filesystem mount point.

Stop IRIS FailSafe and WebFORCE MediaBase on both nodes by entering this command on each node:

# /etc/init.d/failsafe stop

Change /usr/ocs/etc/hostname on both nodes to the new IP alias (the new name used by WebFORCE MediaBase clients).

Edit the file /usr/ocs/etc/hostname on each node and replace the old IP name with the new IP name. For example:

mediabase.company.com

Remember to use the fully qualified IP name.

Edit the file /var/ha/ha.conf on both nodes and make the necessary modifications to replace the old IP name with the new IP name or add the new IP name.

To make the changes take effect, enter these commands on each node:

# /etc/init.d/informix start
# /etc/init.d/mediabase start
# /etc/init.d/mediabase stop
# /etc/init.d/informix stop

Start FailSafe on both nodes and WebFORCE MediaBase on the primary node by entering this command on each node:

# /etc/init.d/failsafe start