The site should be working OK now, if anyone happens to notice problems, could they let me know (a comment on a post, e-mail, Skype, twitter etc.). My hosting company had a “catastrophic” failure with a RAID array (I’d previously seen their storage listed as RAID 5, so that means more than one disk failed at the same time), and it took a while to get the correct backup restored. The first attempt led to a virtually empty folder, with just my other site’s empty folders and a few favicons restored. Unfortunately, it looks as though all of Wednesday’s e-mail were lost, so if you happened to send anything to my email / personal / work AT chrismerriman.com addresses, please do resend them.
A few comments and minor tweaks were lost as the backup used was out of date by a day or so, but I think I’ve repeated the work OK now. I also had a quick check over some of the tags, as DanD noticed that a lot of posts tagged as Dan had nothing to do with him or DanC. It turned out that when I went through a session of mass (manual) tagging a while back, I ended up tagging any post that contained the letters D, A and N in a row with “Dan”. This meant dancing, dancers and sedan would all have triggered an incorrect tag. If anyone happens to notice more incorrect tags (most likely on older posts), again, please do let me know.
I have a LOT of videos I want to embed, as well as months worth of links (mainly from other people’s tweets) that I’ve been meaning to check out and pass along on this site, and obviously I still have a lot of Anna and other photos to catch up with.
Feb 10 2011
February 11th, 2011 12:38 am
More than one disk in an array failed at the same time, ummm, what are the odds 🙂
February 11th, 2011 1:02 am
I think it more likely that if they are running a RAID array, that it is for speed or simply creating a huge array from multiple drives. If there was either mirroring or parity (straight copy/backup or using 1 drive to check contents of two or three others are correct) then one drive going would have resulted in temporary slowdown whilst re-building array on new drive, but to have two (or more) drives go at the same time would mean major issues with either reliability or spikes not being filtered out etc.