You all make use of services that CIS provides, such as email, file storage and web pages. Many of you rely on these services to get your work done. We call the machines that provide these services “production servers”, as opposed to development, test and staging servers. We are going to start having a weekly “maintenance window” for our production servers and network. A maintenance window is a scheduled short period of time during which planned maintenance work takes place. Only emergency work will take place outside the maintenance window, which will provide a more predictable environment for people who rely on the services.
One week before each maintenance window we will circulate information about it via email, the CIS website and postings in the labs. The information will include which systems we will be working on and which services will be affected. This will give you a chance to alert us to any serious scheduling conflicts and to make plans for your own work during the maintenance window. We will also notify you a week in advance if there is not going to be maintenance work in a given week.
At least one month in advance, we will provide notice about planned work that will require a network (wired and wireless) outage.
Our weekly maintenance windows will be Thursdays from 6-9pm and will begin on April 2, 2009. So you can expect information about that maintenance window to be delivered no later than Thursday March 26, 2009.
We will treat the first five weeks of this initiative as a pilot program, during which we will ask for your comments. Your feedback will help us tweak the practice as we move forward.
I wish to thank the HMC Cabinet and the Computing Committee for helpful discussions and advice about this initiative.