Hosting

When I build (mostly Drupal) sites I often do it with the software installers like Softilicious in the control panel. Sometimes they are just to play with, sometimes there are some fatal errors with modules that react badly and I have to start again. The result is that a lot of cruft builds up with databases no longer connected to anything, but taking up space and clutter. Today's lesson could be summed up by the phrase "be careful when clearing the cruft".

There will be some downtime for this site as I have to switch servers. I will try to minimize it as much as I can but it will likely be a few weeks before everything is back in order, bigger and better than before. The hosting plan I'm on is being deprecated, it runs out in about 3 weeks. I am switching the site and the various subdomains to a new and much more flexible plan from the same host, this does mean rebuilding almost everything on the new server, which means a lot of extra work for me to bring things back up to speed.

Yesterday from about 6pm until 3:40am GMT my site was in various states of unreachable. This was due to a fault on the server my webhost owns. I don't know any details other than a reboot around 6pm didn't work. It turned out to be a hardware failure, and everyone's databases etc had to be restored from backups, hence the downtime. This does give me an ideal opportunity to blog about my webhost.

Up until about a week ago I've jumped from one free host to another with mixed results. After prolonged periods of downtime I decided to jump once and for all to a paid host. Free hosting is perfectly workable if you know the limits and are prepared to juggle different accounts for different things. My new host is not only capable of having the site accessible 24 / 7 but gives me much more freedom to really start to work.

As my blog seems to be down as often as it's up over the last couple of weeks I thought I'd explain why. I use FBH as my host. This site is on server 6 (of 6), which seems incapable of lasting a few hours without Apache locking up and requiring an admin to restart it. The problem in the amount of downtime comes from the fact that there is only one admin to do this.