I’ve been toying with the idea of populating my domain and hosting at home. Since I sold my Cobalt Qube I have one machine at home on which to host. It strikes me as a bad idea to host a development site on the system that is also the archive for all of my digital artifacts such as photos, college papers, and correspondence. Of course, I maintain a mostly up-to-date, off-site backup (who doesn’t?), but that doesn’t fully mitigate the inconvenience and possibly disastrous consequences (think keylogger) that might result from an exploited workstation. So, what to do? I could pay someone else for hosting. The pro’s are that it’s relatively inexpensive, there is some expectation of maintained up-time, and my data is not at-risk.

I’ve experimented with virtual machines including Kernel-based Virtual Machine for Linux and VMware Server.  The expectation with a virtual machine is that an exploit in Apache2 would be confined to the virtual server and would not allow access to external file space.  However, virtualization introduces more overhead and, due to increased complexity, increases the likelihood of failure.  That said, no one is paying me to maintain a certain amount of up-time.  I’ve also been thinking of implementing mod_chroot and mod_security, both of which are included in the Ubuntu software repositories.  I have no experience with either, but they seem to be a nice compromise between virtualization and running straight Apache.

The problem I foresee with  running Apache chroot is the difficulty running third-party software in conjunction with Apache.  I’ve been playing with django, dojo, and would like to explore some map applications like Minnesota Map Server.  I also want to work more with Python CGI.  That may make virtualization a simpler environment to configure than chroot.  I’m open to advice.

Looking at the analytics for this site I realized something that should have been obvious even without looking. Every link into this blog, including those provided by search engines, is wrong. People are searching for things like Ubuntu Qube, following the links they find, and landing on a 404 Not Found page. Then they leave. This is bad for visitors and for my page rank.

I don’t really have time to fool with it and my blog isn’t of such compelling interest (e.g., it doesn’t generate revenue) that I feel motivated to do anything about it. However, as a mental exercise I was thinking of how to deal with it. Here’s my plan:

Old links point to things to articles with URL’s in the form of www.prairienet.org/~jtuttle/blog/item/134/.
I could code my 404 page to parse the URL, determine if it’s an old bBlog link, and to obtain the article ID to the old bBlog article. I could then query the bBlog article table for that ID and return the article title. With the article title I could query the new Wordpress article table for that article. If found, I could automatically return the correct article with a 301 Moved Permanently header. I’d have also have to accommodate /section/section_name links. Easy. I’m still not going to do. It’s spring.