There is nothing quite so frustrating as trying to meet a project deadline, when everything is going well, you started on time, are moving at a great pace, and BAAAAAAAAAAAAAAAM!  You hit a brick wall!

That is exactly what happened to me while trying to develop and test a back end server for a project at work.  With heavy traffic, heavy logging, and heavy data, it was essential that I take my application to the brink, stress test, null and boundary test, and just plain out abuse the application as much as I could.

A quick application to spawn UDP traffic was able to provide me with the heavy traffic I needed.  Then the unexpected happened!  I was losing data!  How can I lose data when its all local (within the same servER?!?!)  Well, several days later of researching, googling, diving farther and farther down in to libraries, I found the problem was THE LEAP SECOND!

WHAAAAAAAAAAT!?  Yeah, ugly!  Seems like the Linux kernel has an issue with leap seconds where somehow that ends up affecting server processes.  My symptoms were lost UDP data and general sluggishly slow server.  The solution was simply to stop the ntp server, change the date format, and restart ntp!

Just like that, all my problems were solved, and paradise was achieved.  Mark another loss for time-space anomalies!