The iGEM 2009 wiki was frozen at 11:59pm EDT on October 21. Teams trying to edit the wiki in the hours leading up to the freeze may have experienced some loading issues. Below is an explanation as to what happened.
iGEM HQ had been monitoring the iGEM 2009 wiki all day and all night. It was running very reliably all day and into the evening.
In the last three hours, the load was exceptionally high as teams all over the world modified their sites. The site was running at a wiki page change rate of about one per second. The machine had a load average of 20 to 40. This is on a quad core modern server. The machine did not run out of memory or any other resource.
The wiki did not actually go down, but the database server was at times not able to keep up with the exceptional load and the software said there were technical problems. This problem became significant around 9:30PM and as a result, the server hosting company restarted the mysql server. When the server was restarted, the machine was immediately able to handle pending requests until the point when it once again became overloaded.
In previous years, we have seen a very high load on the wiki on the last day, but the load fell off well before the freeze time. This year there was essentially a giant crowd queuing up requests to the server. The server was able to deal with some number of requests evidenced by the fact that changes were still successfully made at a rate of one per second.
For next year we will work on increasing the power of the servers. However, there is no way to predict how much load the machine will experience in the hours leading up to the wiki freeze. Again, we will suggest that teams manage their time carefully and not expect to do too much at the last minute. We also recommend that teams develop their wiki site on the iGEM server throughout the summer rather than developing the site on another machine and moving it to the iGEM wiki shortly before the freeze.