When technology works, it’s as if it were invisible to the user. That’s why Information Services puts a significant effort into every system implementation or upgrade we do. The August 2 Banner upgrade is a case in point. In order to successfully complete the upgrade in one twenty-four hour period, project managers Tequita Hawkins and Andrea Zinski started work in February. To call it a Banner upgrade is to simplify what was accomplished. In addition to updating the different modules of Banner, this upgrade included the implementation of new database servers, an upgrade to the server operating systems, and an upgrade to the databases themselves. Figuring out the order in which things should be done is painstaking. And it’s not enough to simply plan out each step: real-world technology often creates unforeseen complications that need to be accounted for so the upgrade process doesn’t fail.

Tequita Hawkins and Andrea Zinski work on the Banner upgrade.

That’s why the August 2 Banner upgrade was actually done four times. It began months ago with Walkup Zero, when the team decided what pieces were being upgraded and learned how they interact. Walkup One, the second practice upgrade, was done for documentation purposes. For this upgrade, that meant the creation of an Excel spreadsheet containing more than 80 actions that had to be taken by the final team of ten Systems and Network staffers. Tequita Hawkins, Andrea Zinski, Doug Broome, and Lee Parker represented Enterprise Systems. Vinny Petrone and Sasko Stefanovski were the system administrators for the servers. David Barth, David Foster, and Lennie Rimmer were the database administrators, and Alison Harvey oversaw the network aspects of the final upgrade.

CIO Keith W. "Mac" McIntosh checks in on the upgrade.

After Walkup One was completed, it was time to get the users involved. The upgrade process was done again, in Walkup Two, and it was this version that users in administrative offices across campus were asked to test. User testing began July 1 and ended July 21. Hawkins and Zinski said they were pleased that testing went smoothly and that no major problems were encountered.

That set the stage for Walkup Three, the final production upgrade. It’s not easy to pick a time to upgrade Banner. There’s never a point in the year when at least one office isn’t using the system heavily. The upgrade would mean that Banner and BannerWeb would be unavailable for up to 24 hours as the team went through each upgrade action. That’s a significant amount of time for any office.

Programmers, system administrators, and network specialists work together to upgrade our systems.

Luckily, there is one day that’s slightly easier than the others to take Banner offline: on Busch Gardens Day, the University shuts down at noon so that faculty, staff, and their families can relax and enjoy themselves. Some years, Facilities takes the opportunity of Busch Gardens Day to take the University’s power grid offline for maintenance, but that was not the case this year. And so, at 12:00 PM on August 2, the Banner upgrade team met in the temporary Robins Stadium offices of the system administrators and database administrators to perform the production upgrade. Banner and BannerWeb were taken offline, and all of the systems that interface with Banner had those integrations paused. The upgrade began with a full backup of the production system - a fall-back in case something with the upgrade went wrong, and then continued through the 80-plus actions that had been rehearsed three times before.

Upgrades are complicated, hurry-up-and-wait affairs. At one point, part of the team is hard at work configuring the new system or installing the upgraded software, at another point, everyone waits as the large database is copied from one place to another. The atmosphere is relaxed but focused, as team members wait to do their next actions.

When asked the magnitude of this year’s Banner upgrade, Hawkins and Zinski agreed that this was a “medium-sized” upgrade. Last year’s upgrade was a facelift to the Banner administrative application: it had a significant impact on administrative users on campus. This year’s Banner upgrades were less significant, but the implementation of the new servers, the new operating systems, and an upgraded database made the overall effort more significant for the technical team than for the administrative users.

A little more than twelve hours into the upgrade, the team was wrapping things up. Each step, carefully planned and sequenced had worked, and the newly upgraded system was starting up. Integrations to other systems that depend on Banner were turned back on. The next morning, one user from the Registrar’s Office reported a problem that was soon addressed. After a total of more than 1950 hours of work (not including user testing), the upgrade was complete, and the University was back in business. The technology could fade once again into the background.