Not only has the Blog been upgraded to WordPress version 2.3.3, but it took a total of 5 minutes, with about 30 seconds of downtime and 30 seconds of reconfiguration-while-up time.

Woot. That was was much easier than last time…

We upgraded the WordPress software that runs the blog to the new 2.3.1 version, in response to vulnerabilities present in the old version. There was some downtime this afternoon (intermittent over ~10 minutes) as I switched our installation over to manual (as opposed to Debian package-managed) and then upgraded to the latest version. In the process, I set up a system so that hopefully future upgrades will happen immediately after release of a new stable version, and should cause zero downtime.

Just run some updates on the server. We are now running the latest version of WordPress, Apache2.2 (our http server), and a few other services. Nice to keep up with security upgrades. Also, if you take a look at the URLs of the links, you may notice something different, and if you’re a long-time follower of the site, something familiar. We are now running once more with the PrettyPermalinks option in WordPress, so articles, pages, and comment are linked via the date and title of the posts. Old links still work, such as http://robocup.bowdoin.edu/blog/?page_id=176, but are much more convenient as http://robocup.bowdoin.edu/blog/media/, or http://robocup.bowdoin.edu/blog/2007/08/24/beautiful-aibo/.

The German team have put up a video of our practice match yesterday. Unfortunately, a functioning wireless network we did not have, but we still put up a good fight, ending at 2-1 to the Germans after a half game of play, despite not having a goalie at some points and only having two attackers/defenders.

Our server robocup.bowdoin.edu has undergone a transfer. The virtual image has been moved onto a new host machine, bromlin.bowdoin.edu. This server is running more up-to-date software, seems to be quite a bit faster, and we have had our memory doubled. In all, the server is quite a bit more responsive, and the networking issues we’ve been having have been ironed out. The server can now be rebooted remotely and correct loads all network interfaces and mount the network drives that store our SVN repository, Trac instance, and other data files. Hopefully we will no longer run into any of the outages we have had problems with in the past.

The server handled the switch with no errors, but it did result in network interface identifiers being switched around, which baffled Tucker and myself until I ironed out the network configuration and rebooted several times. On a side note, once up, I updated the virtual machines WMWare tools and Debian/Linux software, running the latest kernel and other patches and updates.

Due to an error on the host machine, this time, we experienced another period of downtime from last night, Feb. 19, to today, Feb. 20. Network connectivity has been restored to the server, and the blog, trac, svn, and media are all running fine after a few reconfigurations. This particular problem is now a reoccuring error on the VMWare server that hosts our machine, and will hopefully be resolved, as it affects not only our own server but several others as well. Hopefully no one was delayed from working to long last night, but all should be in working order today.

Our blog had a short amount of down time recently, due to an error on the server. To get technical, a few uncleaned caches resulted in the /var partition filling up, leaving the mysql server without enough space to operate. Thus, the blog, which using WordPress on Apache2 and stores it’s data in a MySQL database, could not load this data. The offending caches have been removed, and the mysql and apache services restarted. Everything is now running normally. A few changes should result in this not happening again.