I’ve been doing some prep-work for our new Blackboard Learn environment that will be launched here at Palomar for Spring 2011 (knock wood). Part of that is system documentation, and there’s a massive amount of procedural data that is included along with information like serial numbers and IP addresses. Working through the operations workbook for our existing Blackboard 8 system I looked at some of our times at which we consider it acceptible to have the system down for maintenance:
There are daily and weekly windows when the system can be routinely backed up, where sluggish response times are acceptable. A short (up to an hour) window of system down-time is acceptable on the “Reboot Wednesday” times between 6 and 7 a.m. on the first Wednesday of each month. However, the only time when extended system down-time is acceptable would be during the annual windows.
Daily: 2 a.m. to 7 a.m.
Weekly: Saturday, 7 a.m. to 12 p.m.
Monthly: First Wednesday, 6 a.m. to 7 a.m.
Annually: The first full week in January, also any time window between the end of the Summer term and beginning of Fall.
Generally speaking we don’t use the Daily time windows, or Weekly ones, but we do try to restart all the servers on Reboot Wednesdays. (This deals with several possible technical problems that may bring the system down unexpectedly. If it didn’t help things, believe me, we wouldn’t do it.) The support line and ticketing system seldom see any activity during these times, which leads me to believe that few people are impacted by those downtimes.
Whenever we do upgrades and significant maintenance to the Blackboard environment between semesters, we do receive quite a number of calls… but there really isn’t any way around that, sadly.
Obviously there are other times when the system may seize up, and it’s inaccessible to users. We try to get things running again as quickly as possible, of course, but that sort of thing isn’t what we schedule for. It is possible that if we do schedule more frequent maintenance windows, akin to what we do on Reboot Wednesdays, that we might see fewer or no other system seizures, though.
The reason I’m sharing all this is that I’m interested how the plans we have for Palomar measure up. So, for Palomar faculty, please comment on this post and let me know how you feel about the idea of having more frequent, planned, system outages. For non-Palomar faculty, please comment about how this sort of maintenance schedule measures up against what you’re accustomed to at your institutions. I’d really appreciate hearing what everyone thinks on this topic!