Author Topic: Move over Y2K "computer chaos" feared by the Leap second  (Read 1837 times)

0 Members and 1 Guest are viewing this topic.

Offline HighPockets

  • Frequent Poster
  • ******
  • Posts: 2278
    • View Profile
Move over Y2K "computer chaos" feared by the Leap second
« on: January 09, 2015, 08:06:53 PM »
What are you all going to do with all the extra time we get this year?

http://www.usatoday.com/story/tech/2015/01/08/computer-chaos-feares/21433363/
Everyone you will ever meet knows something you don't.

Offline PANTS!

  • One leg at a time.
  • Poster of Extraordinary Magnitude
  • **********
  • Posts: 12181
  • What seals I auditioned for this job.
    • View Profile
Re: Move over Y2K "computer chaos" feared by the Leap second
« Reply #1 on: January 09, 2015, 09:24:15 PM »
One leap forward
Two leaps back
Will politics give me the sack

Waiting for the great leap forward.
« Last Edit: January 10, 2015, 12:24:13 PM by PANTS! »
Large data of text  or Mediumtext or Longtext

Offline The Latinist

  • Cyber Greasemonkey
  • Technical Administrator
  • Reef Tank Owner
  • *****
  • Posts: 8876
  • Fanatically anti-bigot
    • View Profile
Re: Move over Y2K "computer chaos" feared by the Leap second
« Reply #2 on: January 10, 2015, 11:27:32 AM »
Yeah, leap seconds are stupid.
Large data of text  or Mediumtext or Longtext

Offline SaladTimeAgain

  • Brand New
  • Posts: 4
    • View Profile
Re: Move over Y2K "computer chaos" feared by the Leap second
« Reply #3 on: January 12, 2015, 01:43:29 AM »
What are you all going to do with all the extra time we get this year?

I will eat one bite of a burrito

Offline syameese

  • Off to a Start
  • *
  • Posts: 39
    • View Profile
Re: Move over Y2K "computer chaos" feared by the Leap second
« Reply #4 on: January 12, 2015, 04:19:24 AM »
Well, if leap second fear gives us a movie as good as Office Space I am all for it  ;D

Offline Pusher Robot

  • Frequent Poster
  • ******
  • Posts: 2217
  • Do you have stairs in your house
    • View Profile
Re: Move over Y2K "computer chaos" feared by the Leap second
« Reply #5 on: January 12, 2015, 01:31:08 PM »
I just don't see how this will have such big impact.  Programs using actual dates and times, or Unix epoch, must already have been debugged to deal with shifts due to hardware clock drift corrections.  Programs requiring more precise timing already use a system timer that isn't tied directly to UTC, like number of milliseconds since previous midnight or system uptime or something.
Large data of text  or Mediumtext or Longtext

Offline The Latinist

  • Cyber Greasemonkey
  • Technical Administrator
  • Reef Tank Owner
  • *****
  • Posts: 8876
  • Fanatically anti-bigot
    • View Profile
Re: Move over Y2K "computer chaos" feared by the Leap second
« Reply #6 on: January 12, 2015, 03:26:42 PM »
I just don't see how this will have such big impact.

I read that in 2012 the bug took down sites including Reddit, Foursquare, Yelp, LinkedIn, Gawker, and StumbleUpon.  Gawker's servers became completely unresponsive and and could not be remotely reset; they had to actually physically reboot their servers.  I also read that a lot of the issues were actually caused by Java and affected applications written in java, including certain database software.
Large data of text  or Mediumtext or Longtext

Offline spawnstar

  • Not Enough Spare Time
  • **
  • Posts: 219
    • View Profile
Re: Move over Y2K "computer chaos" feared by the Leap second
« Reply #7 on: January 12, 2015, 05:14:13 PM »
http://en.wikipedia.org/wiki/Year_2038_problem

This could be a problem. It would only cause problems with 32-bit systems though.

Offline arthwollipot

  • Reef Tank Owner
  • *********
  • Posts: 9780
  • Observer of Phenomena. Pronouns: he/him
    • View Profile
Re: Move over Y2K "computer chaos" feared by the Leap second
« Reply #8 on: January 12, 2015, 06:20:24 PM »
Since most timekeeping devices these days automatically update, how can this be a problem? My iPhone already automatically adjusts for daylight saving and timezone shifts. If there hadn't been such hype about it, I probably wouldn't even know the leap second existed.

Y2K on the other hand was actually a thing. Many of the claims about its effects were vastly exaggerated, and the main reason nothing happened is that armies of people worked very hard to make sure nothing happened, but it was a thing to be concerned about. And in fact nothing didn't happen - there were plenty of reports following Y2K of computer systems spitting out the wrong date.
Self-described nerd.

Offline The Latinist

  • Cyber Greasemonkey
  • Technical Administrator
  • Reef Tank Owner
  • *****
  • Posts: 8876
  • Fanatically anti-bigot
    • View Profile
Re: Move over Y2K "computer chaos" feared by the Leap second
« Reply #9 on: January 12, 2015, 07:10:58 PM »
Since most timekeeping devices these days automatically update, how can this be a problem?

The issue is with computers using the Network Time Protocol for time synchronization rather than using a local system clock.  Such interconnected systems need to keep I/O synchronized across multiple machines, sometimes located far from each other.  They do this by reference to a time server connected to an atomic clock which creates a continuous time signal that these computers can use for synchronization.

During a Leap Second, the NTP server does not follow the usual pattern of 23:59:59 --> 00:00:00 but instead sends the time signals 23:59:59 --> 23:59:60 --> 00:00:00.  If an application or system software is not set up to handle these time signals properly, then it may fail to keep operations properly ordered.  How that affects the system depends on how robust the rest of the software design is and how important the offending software is.
Large data of text  or Mediumtext or Longtext

Offline arthwollipot

  • Reef Tank Owner
  • *********
  • Posts: 9780
  • Observer of Phenomena. Pronouns: he/him
    • View Profile
Re: Move over Y2K "computer chaos" feared by the Leap second
« Reply #10 on: January 12, 2015, 07:14:16 PM »
During a Leap Second, the NTP server does not follow the usual pattern of 23:59:59 --> 00:00:00 but instead sends the time signals 23:59:59 --> 23:59:60 --> 00:00:00.  If an application or system software is not set up to handle these time signals properly, then it may fail to keep operations properly ordered.  How that affects the system depends on how robust the rest of the software design is and how important the offending software is.

Thank you for that explanation - it makes perfect sense.
Self-described nerd.