Q. Can ITS employees who perform scheduled maintenance avoid having to come in every Thursday at 5:00 AM?
A. Yes, they can. In fact, no one will ever have to come in early every Thursday, and most people who use the maintenance window to perform updates and apply patches will not come in early more than a handful of times during the year.

Here are some options that can be leveraged to avoid having to be here early most Thursdays:

• Perform non-disruptive maintenance during regular hours. If work can be done to the systems that doesn’t interrupt access to campus services, then that work doesn’t need to be performed during the maintenance window. As a rule of thumb, I would consider work non-disruptive that doesn’t impair a use of a IT-provided service for longer than 1 minute.
• Don’t perform any disruptive maintenance that week. If there is no need to perform disruptive maintenance, then no one needs to come in early.
• Batch maintenance tasks together. Skip a week or two or three between performing routine maintenance, batching up patches and upgrades into larger bundles. However, this needs to be balanced with the more important goals of not falling too far behind in applying critical updates, not expanding the disruption period beyond the 2-hour window, and the increased difficulty of troubleshooting or rollback when things don’t work out as expected.
• Rotate the maintenance duties. Instead of the same person always doing the patching and updates, two or three people rotate. This has the added bonus of ensuring knowledge sharing on our critical systems.
• Schedule server reboots. When a server needs to be rebooted to fix a memory issue, apply a patch or perform a minor update, it can be scheduled to auto-reboot during the maintenance window. If there is doubt that the server will come up normally after the reboot, the first one or two times the work is done, the reboot can be monitored from either on campus or remotely from home. When a comfort level with the reboot process has been reached, then the reboots simply become part of normal operations and are monitored via WhatsUp, alerting us if it does not come back up at the expected time.
• Automate. Use scripting and other management tools to automate maintenance processes so that the work can be done during the window without anyone actually having to trigger it.

So these are a few ideas. There are probably some other good ones just waiting to be discovered. Again, the key concept of the maintenance window is that the time is reserved for us to use when we need it—not that we have to use it each time—and that disruptive maintenance takes place at a time when the campus expects it and it is minimally disruptive to the work of the campus community.

Q. How should time spent on work during the maintenance window be tracked?
A. If someone performs maintenance during the weekly time window either on campus or from home, they can balance that out by taking off earlier either the same day or another day within the same week, such that the total weekly time worked still comes out to 40 hours. It should not be considered overtime or comp time, just a temporary adjustment of their regular weekly work schedule.

Q. Will there be an evaluation period?
A. At least annually, we will review our maintenance window policy to see if there are ways to improve on it.