Server Outage December 20, 2009 - January 3, 2010
Description
As a result of an ongoing issue with Blackboard Vista's garbage collection mechanism, we are going to take Blackboard Vista offline between December 20, 2009 and January 3, 2010. Much of this time the University is officially closed for winter break, though we realize users may be impacted. Unfortunately, this longer than usual maintenance outage is necessary in order for us to ensure that the Blackboard Vista system remains online and reliable through June 2011, at which time our support license ends for this product. We will be contacting users with migration information throughout 2010, as we move courses off of the Blackboard Vista server. To keep informed about LMS updates, including migration plans as they become available, see http://wikis.lib.ncsu.edu/index.php/LMS_Strategy.
If you have questions, comments or concerns, please let us know as soon as possible. You can contact us via the LearnTech help desk, at learntech@ncsu.edu.
Thank you for your patience and understanding,
DELTA
Current Status
The current status of the outage will be posted to http://sysnews.ncsu.edu.
Technical Details
Our Blackboard Vista server has been unable to run Database Garbage Collection (GC) for more than 24 months. Normally, GC runs continuously and manages the small amount of garbage (e.g. deleted files) with little effort. Our GC process has not been running correctly since at least March of 2007. The database is currently approximately 1.3 terrabytes in size, 70% of which is garbage.
All that garbage makes backups, maintenance, and everything we do take much longer than it should, and has us concerned about Vista's ability to optimally run through 2011 (when the license expires). We have worked on this issue for a long time with WebCT support, then Blackboard support, as this has been a known problem for some time. After many frustrating months working with their support, we have a solution that we have verified in our test environment, but which will take time to implement on the production server.
With the window of downtime over the Holidays, we will:
- Turn off access to Vista
- Copy our production database to another database server (approximately 2 days)
- Run the manual GC process we have verified in our test environment (approximately 14-20 days). GC will run much faster with no users on the system, and it should be finished or be very close to finished by the time we restore access on January 3rd.
- Restore access to Vista on 1/3/2010, whether GC is finished or not. (Spring classes begin 1/11/2010)
- Reboot the entire cluster on 1/17/2009 and restart automatic GC.