Michael Doran Home Page
Contact | Site Map | Search  
This page is deprecated: please read archives disclaimer.

Implementing a
Stand-Alone Continuous OPAC

Introduction | Overview | Tasks

Important Note
This tutorial is for system administrators with significant Solaris and Voyager experience. Some steps are described in general terms with the expectation that the sysadmin will be able to fill in the details and also understands the ramifications involved. Care should be exercised when editing Unix system files and restoring data from backups.


Acronyms
The following acronyms will be used in order to avoid repeating these two mouthfuls.

VDBS - Voyager Database Server
SCOPAC - Stand-alone Continuous OPAC Server

Requirements
  • Separate server (e.g. application server) to serve as SCOPAC
    • Same operating system as VDBS (e.g. Solaris)
    • Sufficient disk space (preferably a dedicated disk)
    • Apache web server software installed and configured
Advance Preparations
Familiarize yourself with the concept of machine aliases and how the Domain Name Service (DNS) operates. Find out who administers DNS at your site and how to go about editing entries.

Create all the relevant Voyager users and groups on the SCOPAC. Copy over each user's .profile file from the VDBS. The "library" user's home directory will also require a symbolic link to the voyager.env file.

Prepare a disk
The Voyager files must reside on one file system. It is highly recommended that you dedicate a disk to this purpose. Our database of approximately 1,000,000 records fits comfortably on a 35GB disk, and could probably be shoe-horned onto a 20GB disk. In calculating your disk requirements, keep in mind that the /m1 filesystem tends to get bloated over time with non-essential files. Not everything is needed for a SCOPAC instance.

Install Voyager database on SCOPAC
Schematic illustrating the transfer of VDBS files to SCOPAC There is more than one way to get the needed files where you want them. I recommend restoring the required Voyager files from your most recent VDBS backup to the SCOPAC disk. Doing so not only accomplishes the task at hand, it also gives you real-world, verifiable practice restoring from backup.

Configuration on SCOPAC
  • Add Voyager entries to the /etc/inet/inetd.conf file.

    The SCOPAC /etc/inet/inetd.conf file needs to have the same Voyager Production Database entries as the inetd.conf file on the VDBS. (Suggestion: Keep two versions of that file on the SCOPAC: inetd.conf.orig and inetd.conf.scopac. Use the one which is appropriate for the task at hand.)

  • Add Voyager entries to the /etc/inet/services file.

  • Create symbolic link:   ln -s /m1/oracle /oracle

  • Edit the Oracle network config files listener.ora and tnsnames.ora so that they use the hostname of the SCOPAC rather than the VDBS.

  • Transfer over the Oracle oratab file.

  • Transfer over the Voyager and Oracle startup scripts: /etc/init.d/dbora and /etc/init.d/voyager. These scripts may or may not require some editing. Use your judgement as to whether it is worth the effort to create links for the various runtime levels. If you don't, just remember that you will have to start and stop Oracle/Voyager manually via the command line.

  • Transfer over the /usr/local/bin/oraenv and dbhome files. These will not need to be edited, but like oratab should retain the original file ownerships and permissions.

  • If necessary, edit the Oracle parameter file, initLIBR.ora.

  • Add the Oracle memory settings to the /etc/system file.

  • Edit any Voyager config files (i.e. opac.ini and voyager.ini) that contain the hostname. This may not be needed if you are using an alias in these files.

Merging files from a separate WebVoyage server
Schematic illustrating the transfer of WebVoyage files to SCOPAC If you maintain a separate WebVoyage server, then the VDBS WebVoyage configuration files may not be the ones you want to use. If this is the case, you would want to copy over the WebVoyage configuration files from the WebVoyager server to the SCOPAC.

Ensuring read-only access
As the name suggests, the main purpose of Continuous OPAC is to allow access to your catalog while the production database is being upgraded. Patrons (and staff) will see a static "snapshot" of the database at a particular point in time. Since this snapshot will never be reconciled with your production database, you don't want to allow any updates to the SCOPAC database.

In addition to OPAC access, you may also choose to allow staff access via the Voyager client modules. This may determine what read-only access method you choose.

Testing
You definitely want to test SCOPAC functionality before it is actually needed. As you start up oracle and voyager (/etc/init.d/dbora start and /etc/init.d/voyager start), you may get error messages. These error messages will usually give you a clue as to what part of your scopac installation needs attention. Chances are good that it is a permissions problem.

Because WebVoyage is designed to easily access an Oracle database on another server, it is sometimes difficult to determine just where it is getting its data. A good test is to point one of your staff clients (e.g. the Cataloging module) at the SCOPAC (by editing the voyager.ini on your PC), make a change to a record, and then verify that the change shows up in your SCOPAC WebVoyage instance and not in your production WebVoyage instance.

It is recommended that you do an initial proof-of-concept of SCOPAC prior to any Voyager upgrade. Once you are confident in your SCOPAC setup skill, timing it for an upgrade is a balance between using the most current Voyager database backup for your snapshot vs. leaving enough time for last minute verification and testing.

Switching over
If you use an alias for your WebVoyage server, you simply have to delete the alias from the DNS entry for your production WebVoyage server and add it to the DNS entry for your SCOPAC server.

It is also a good idea to add a short blurb to the WebVoyage header file (or a link to page explaining about the upgrade) so that users are alerted to the fact that the data may be stale.