The Burning Man SysAdmin Team is responsible for maintaining the technological infrastructure that supports the year-round operations of the Burning Man Project. The web and mail services that we provide can be thought of as Burning Man’s nervous system, delivering information to and feedback from the remote reaches of our participant community.
Over the past year, the SysAdmin Team has continued its tradition of bringing steady improvements to this all-important infrastructure. Here are some highlights from our activities:
We once again moved our web and mail servers to a new co-location facility. We hadn’t planned on this move, but of the two providers that we used to have, one went out of business and the other had recurring service problems. We did some research into various providers in the area and ended up renting a cage space from United Layer. We’ve been very pleased with the service and support that we’ve received from them, and they were gracious enough to remove our bandwidth cap during the course of the event to support our web cast.
The roll-out of the Black Rock City Extranet in April of 2003 was one of the most exciting Burning Man Tech Team developments of the past year. The SysAdmin Team purchased and deployed a new server for this purpose. We experienced sporadic troubles with this machine, which have been tracked to a faulty IDE RAID controller. We’re currently configuring another machine to serve the Extranet for a while so we can replace the current machine’s RAID controller with a higher quality card.
We also installed three other new servers last year. One is currently acting as an office intranet server, another is a firewall, and still another was used to provide database and web services for the Rangers on the playa.
One big success of the past year was the installation and deployment of the SpamAssassin mail filtering application on our mail server. Burning Man has had clear text email addresses on its website for years, and as a result nearly everyone who has a “burningman.com” email address sees a flood of junk mail every day. Getting SpamAssassin to work on a mail server that is as heavily worked as ours proved to be much more difficult than expected. However, we did manage to get it working, and it has done much to reduce the amount of time wasted hitting the “delete” button.
Countless smaller projects have been successfully accomplished. These include installation of a Nagios server monitoring system, installation of a UPS auto-shutdown mechanism for our office servers in case of extended power outage, improved backup systems (although there’s still much to be done here), improved push scripts for use by the Web Team, and many more.
Finally, a fair amount of work went into improving the quality of the internet connection to our San Francisco office. We had been reasonably happy with the SDSL connection we were using, but when our service provider went out of business we discovered that none of the other providers could offer us the same performance. We have since installed a T1 connection from our office to our server cage at the co-location center. We also have become a node in the SFLan project, a San Francisco city-wide wireless network. This allows us to have a redundant connection to our office while extending the reach of SFLan’s coverage.
Submitted by,
Rob Miller
Tech Department Lead