The Channel Islands recently suffered from a massive power failure. Well, in fact, it was 2.5 power failures in the space of three days (one on the Monday, again on Wednesday around 7pm, then again at about 9pm, about 5 minutes after the power had been restored.)
One of my clients suffered from a fairly extensive failure as a result – a couple of disks in their SAN were lost. This had serious consequences for the SQL Server that was hosting some SharePoint databases on top of it. We experienced lots of page corruptions, and, long story short, the SharePoint content database was left in a mangled state. And to make matters worse, in a “perfect storm” of both independent sets of backups from the previous night were also lost. (How this could have been allowed to happen is a topic for another day.) After exhausting all known options, we were left with a situation where we had a backup which was about 2 weeks old and half a database from which to try and recover some of the recent data. We then broke the biggest SharePoint rule: never go anywhere near the SharePoint database.