Why this year's leap second crashed more systems?
According to Wikipedia (http://en.wikipedia.org/wiki/File:Leapsecond.ut1-utc.svg) , this is not the first year leap second is being inserted. Why so many systems crashed now, but not in, say, 2009?
Date/time handling is notoriously difficult to get right, and leap years are an edge case that's triggered very rarely, so it's easy for novice developers to fail to handle it.
[source: http://www.mail-archive.com/git-commits-head@vger.kernel.org...]
I know that CentOS/RHEL 6 was affected, but CentOS/RHEL 5.5 (kernel 2.6.18) was not.
So I guess it's just a coincidence, a bug was introduced in a code path that's rarely executed and so lay dormant for some time, gaining more and more victims as time went by and people adopted kernels containing said bug.