Wow, November rolled around way too quickly this year, didn't it? Wasn't it just summer? And where the hell did autumn go? Anyways, with the winter, and imminent fuckening approaching, the tinkering and shenanigans must go on. I haven't done anything with the gopher server, nor content since the last post, but I did decide to attempt something as mundane as changing the IP range of the server cluster this site, among other things are hosted on. Shouldn't be too hard right? Hah! Nope! Well, if everything had gone to plan, it would have been pretty easy. The tricky bit is that there are several servers connected together as a cluster, with some- thing called corosync running between them to keep everything synchronized and in check. This means that you, in order to change the IP series, or address at all for that matter, need to edit serveral different config files, and make sure to up- date the version number in the corosync config. None of that is tricky, if it wasn't for the fact that the corosync inter- faces already had the same IP range as I wanted to use.. So I changed all the things around accordingly, stuff broke, as to be expected.. But stuff never un-broke itself as it should have. And to add to the fun, I could now no longer edit the corosync config anymore. I somehow didn't have permissions to do so. So I now had a completely broken server cluster. Wooh! What now, then? Well, I still had one node still up and runn- ing thankfully. So I rather nervously made backups of all the virtual machines running on that node. I then went on the fun adventure of completely reinstalling and rebuilding the clus- ter from scratch. This all went well without any issues. But the fun then began again when I set out to restore the virtual machines from backup.. I got all sorts of "funny" non-descript "io-error" warnings and rebuild processes that decided to go all the way to 100% and then just sit there for hours, just being stuck. This turned out to mostly be due to me being impatient and trying to restore too many machines at once, causing the cluster nodes to freak out slightly.. So I reversed a few steps and tried again with a bit more patience. And as you can see, it all went well in the end, as this page is hosted on said cluster. :) Anywhoos.. That's enough blabber for now I think. Cheerio! -Headcrash