If you can read this, either the style sheet didn't load or you have an older browser that doesn't support style sheets. Try clearing your browser cache and refreshing the page.

(Some Pirate)   Millions of dollars and years of lobbying by the RIAA and all it took was a leap second to sink The Pirate Bay   (zeropaid.com) divider line 45
    More: Fail, The Pirate Bay, RIAA, leap seconds, social news  
•       •       •

10034 clicks; posted to Geek » on 01 Jul 2012 at 12:27 PM (2 years ago)   |  Favorite    |   share:  Share on Twitter share via Email Share on Facebook   more»



45 Comments   (+0 »)
   
View Voting Results: Smartest and Funniest

Archived thread
 
2012-07-01 12:19:03 PM  
From the quote: TPB will be back as soon as we can get a sober admin to fix the problem.

I see Fark and The Pirate Bay enjoy the same sysadmin philosophy
 
2012-07-01 12:38:12 PM  
So, Reddit, Fark, TPB, and b went down last night? BAH GAWD, Mike was behind it all wasn't he. He wanted to force us to watch the ustream as our only source of entertainment. He's an evil computer-based supergenius, don't worry, I have a solution to save us all. Everyone together, let's #######carrierlost###########
 
2012-07-01 12:40:05 PM  

Snarcoleptic_Hoosier: I see Fark and The Pirate Bay enjoy the same sysadmin philosophy


If it works, don't fix it. :)
 
2012-07-01 12:46:22 PM  
So a bug that far as I can tell was discovered back in 2009 is still able to take down linux based servers today? Good jorb guyz!
 
2012-07-01 12:50:10 PM  

TheManofPA: So, Reddit, Fark, TPB, and b went down last night? BAH GAWD, Mike was behind it all wasn't he. He wanted to force us to watch the ustream as our only source of entertainment. He's an evil computer-based supergenius, don't worry, I have a solution to save us all. Everyone together, let's #######carrierlost###########


Even facebook was assholed for a while. Had it been during the middle of a workday it would have went down as the most productive day in recent history.
 
2012-07-01 12:52:19 PM  
Wow, if just a second can cause that type of carnage, imagine how bad that Y2K bug is going to be...
 
2012-07-01 01:06:42 PM  
I wonder if this could be used as a DoS attack?

Send a spoofed NTP packet with a value of 23:59:60 to a Debian Linux server.
 
2012-07-01 01:16:34 PM  
Thank goodness I'm on the deep web.
 
2012-07-01 01:20:28 PM  
This DEFINITELY heralds the year leap second of linux on the desktop.

Every time your computer locks up completely we shouldn't call it a crash. Crashes are for hard drives. These sorts of failures should be called a "Linux". An error that brings an entire system to it's knees because of a bug known to every single user, available to be fixed by anyone, that nobody did anything about.

Knowing there is an issue and doing farkall about it because uptime/i'm a guru/won't happen to me/ I'm drunk. Is the quincentennial linux way.
 
2012-07-01 01:23:28 PM  

count chocula: I wonder if this could be used as a DoS attack?

Send a spoofed NTP packet with a value of 23:59:60 to a Debian Linux server.


Yup, it certainly seems if you can pretend to be a responsible NTP service you can potentially knock the thing down. Considering it was discovered in 2009 but Redhat only patched it in late April of 2012 I'd say the odds are pretty good. Its a classic deadlock that works best on machines that are relatively active. The heart of the matter is during that second a printk is called to write into the logs that the leap second came into play but if something else happens to printk before its done then they just lock up and wait for one another forever.
 
2012-07-01 01:44:28 PM  

fluffy2097: This DEFINITELY heralds the year leap second of linux on the desktop.

Every time your computer locks up completely we shouldn't call it a crash. Crashes are for hard drives. These sorts of failures should be called a "Linux". An error that brings an entire system to it's knees because of a bug known to every single user, available to be fixed by anyone, that nobody did anything about.

Knowing there is an issue and doing farkall about it because uptime/i'm a guru/won't happen to me/ I'm drunk. Is the quincentennial linux way.


I sure hope you win your good fight against those evil Linux bastards, may they burn in hell for all the misery they've caused.
 
2012-07-01 01:48:14 PM  

KellyLockhart: Snarcoleptic_Hoosier: I see Fark and The Pirate Bay enjoy the same sysadmin philosophy

If it works, don't fix it. :)


Probably why I frequent both.
 
2012-07-01 01:58:28 PM  
HA! Stupid geek nerd fanboys. brought down by a leap second. If they used a REAL operating system--I'm talking Windows here-- there wouldn't have been any problem.

You get what you pay for, losers!

Thank God your kind will eventually go extinct since all of you Linux pukes are going to die virgins.
 
2012-07-01 02:13:50 PM  

MusicMakeMyHeadPound: I sure hope you win your good fight against those evil Linux bastards, may they burn in hell for all the misery they've caused.


Linux is great.

It's the admins and programmers that are the problem.

a bug from 2009 that still exists in OPEN SOURCE SOFTWARE?

and this bugged code is still running on production servers 3 years later?
 
2012-07-01 02:32:32 PM  
A leap second has been a regular occurrence since 1972, there has been 25 in the last 40 years. Systems were not ready, why?
 
2012-07-01 02:35:05 PM  

MrBentor: A leap second has been a regular occurrence since 1972, there has been 25 in the last 40 years. Systems were not ready, why?


Because Linus Torvalds is busy calling Nvidia motherfarkers to fix is own system clock.
 
2012-07-01 02:39:06 PM  

fluffy2097: MusicMakeMyHeadPound: I sure hope you win your good fight against those evil Linux bastards, may they burn in hell for all the misery they've caused.

Linux is great.

It's the admins and programmers that are the problem.

a bug from 2009 that still exists in OPEN SOURCE SOFTWARE?

and this bugged code is still running on production servers 3 years later?


Which probably wasn't even the issue.

On a tangent, I dropped a production database table this weekend. It was a margarita-related error, as that's what I was enjoying when i executed the copy-pasted sql before proofreading it like the dummy I am. I made a flawless recovery but if the boss finds out, I'm totally blaming the leap second.

My point is that admins are human too, you know and we slightly resent having to work seven days a week.
 
2012-07-01 02:44:08 PM  

MusicMakeMyHeadPound: My point is that admins are human too, you know and we slightly resent having to work seven days a week.


If you are an admin that's always on-call, you've done it so very wrong.

Does your boss know you're drunk on company time?
 
2012-07-01 02:55:49 PM  
Do Linux servers keep in constant contact with time servers? Or are they all programmed to check at 00:00:00?
 
2012-07-01 02:57:17 PM  

fluffy2097: If you are an admin that's always on-call, you've done it so very wrong.


There's a difference between being "on-call" and being "called on". I'm technically on call 24/7, but I haven't actually had to go in outside normal business hours for anything unplanned in a very long time. I think the last time was well over a year ago when I started getting weird temperature problem reports from a PBX and couldn't figure out why remotely.

/ turned out the probe was failing
 
2012-07-01 02:57:24 PM  

fluffy2097: MusicMakeMyHeadPound: My point is that admins are human too, you know and we slightly resent having to work seven days a week.

If you are an admin that's always on-call, you've done it so very wrong.

Does your boss know you're drunk on company time?


If you're working sober, you've done it so very wrong.

Jesus, kids these days.
 
ZAZ [TotalFark]
2012-07-01 03:14:21 PM  
If it works, don't fix it. :)

The students programmers at MIT Project Athena in the late 1980s were called "watchmakers" after the alien creatures from The Mote In God's Eye who couldn't resist fixing things that weren't broken. The sysadmins were "droogs," after A Clockwork Orange.
 
2012-07-01 03:48:58 PM  

MarkEC: Do Linux servers keep in constant contact with time servers? Or are they all programmed to check at 00:00:00?


Standard NTP starts querying network-based time servers every 64 seconds and eventually increases to make queries 128, 256, 512, and 1024 seconds as it learns the local clock's properties and can stabilize it better between queries.

In my case, I have a GPS clock that sends the current time (to the second) to NTP. NTP checks that clock every 16 seconds. It also emits one pulse per second to the system, again to NTP. This pulse is accurate to within one microsecond of UTC, so the system clock is always right on.

NTP alerts the system it's running on that a leap second is pending soon (that day, if I recall correctly).

Honestly, timing-related bugs should really be worked out: NTP is one of the older protocols still in active use on the internet and the kernel/other software really shouldn't have any sort of timing-related bugs (particularly for common things like leap seconds).
 
2012-07-01 04:10:09 PM  

Riche: HA! Stupid geek nerd fanboys. brought down by a leap second. If they used a REAL operating system--I'm talking Windows here-- there wouldn't have been any problem.

You get what you pay for, losers!


The truth is that if a bunch of domain controllers (and by extension, entire domains since they'd all bug out at once in this case) was brought down by a leap second, Microsoft would have it's feet held to the fire to answer some serious questions and it would be fixed very quickly.

Linux has nobody to hold accountable for this sort of thing, and thus no drive to fix existing bugs.
 
2012-07-01 04:14:13 PM  

fluffy2097: Riche: HA! Stupid geek nerd fanboys. brought down by a leap second. If they used a REAL operating system--I'm talking Windows here-- there wouldn't have been any problem.

You get what you pay for, losers!


The truth is that if a bunch of domain controllers (and by extension, entire domains since they'd all bug out at once in this case) was brought down by a leap second, Microsoft would have it's feet held to the fire to answer some serious questions and it would be fixed very quickly.

Linux has nobody to hold accountable for this sort of thing, and thus no drive to fix existing bugs.


Yah in the actual bug report for this the guy that I guess is in charge of determining severity of the bug originally blew it off as a none issue that's rare enough to just ignore. I bet he's feeling mighty sheepish today.

Link to the bug report from back in 2009 via redhat
 
2012-07-01 04:22:56 PM  

fluffy2097: Linux has nobody to hold accountable for this sort of thing, and thus no drive to fix existing bugs.


You mean a bug that was fixed in all Linux kernels after version 2.6.29. The latest stable release is 3.4.4. Now, sure, their version numbering scheme changed, but as of 2.6, they've been on a roughly quarterly release schedule. The last version of 2.6 was 2.6.39- which means there were ten releases through 2009-2011 that were not impacted by the bug.

So, remind me- who has no drive to fix existing bugs? The software developers or the user who doesn't keep their system patched?
 
2012-07-01 04:26:25 PM  
fluffy2097
Linux has nobody to hold accountable for this sort of thing, and thus no drive to fix existing bugs.

Not to get in the way of crushing those good-for-nothing free software jerks (what assholes!), I honestly can't find evidence of this being a Debian fault.

The only people reporting that it was a kernel issue is one journalist (whose article this thread is about) based on their hasty interpretation of a forum they read and an incident from three years ago, and a few echo bloggers.
 
2012-07-01 04:43:00 PM  

fluffy2097: Riche: HA! Stupid geek nerd fanboys. brought down by a leap second. If they used a REAL operating system--I'm talking Windows here-- there wouldn't have been any problem.

You get what you pay for, losers!


The truth is that if a bunch of domain controllers (and by extension, entire domains since they'd all bug out at once in this case) was brought down by a leap second, Microsoft would have it's feet held to the fire to answer some serious questions and it would be fixed very quickly.

Linux has nobody to hold accountable for this sort of thing, and thus no drive to fix existing bugs.


Right, no one ever fixes existing bugs in Linux and they feel no reason to... Changelogs are just made up on the spot. Do you realize how ridiculous you sound?
 
2012-07-01 04:56:15 PM  
Response from original bug report.

>The simplest solution for the older RHEL kernels is probably just to remove the
>leap second printks.

Thats some fine coding there lou.
 
2012-07-01 05:14:32 PM  
This is why I patched my system to be compliant with cubic time.
 
2012-07-01 05:21:34 PM  
I wonder if all these huge earthquakes over the past decade are going to effect stuff like this more often. I seem to recall reading that the earthquake that caused the tsunami devastating Japan last year made the year something like 4 seconds longer. The the Boxing Day tsunami from a few years back apparently had an even bigger impact.
 
2012-07-01 05:25:53 PM  

bobbette: This is why I patched my system to be compliant with cubic time.


But have your patched your system to be compliant with four simultaneous days? Otherwise you will have egg on your face when Time Cube time is (finally) accepted.
 
2012-07-01 05:45:12 PM  
What was the reason for the time change, and will we lose an hour and one second when we turn off daylight saving time?

/really? over a single second? you know who's not getting laid much...
//if it's only 1, why call it 'second'?!
 
ZAZ [TotalFark]
2012-07-01 05:59:49 PM  
What was the reason for the time change

It keeps mean solar time aligned with the time zone at the Greenwich Meridian. If we never added leap seconds the sun would be rising 30 seconds late.
 
2012-07-01 06:58:55 PM  

fluffy2097: Response from original bug report.

>The simplest solution for the older RHEL kernels is probably just to remove the
>leap second printks.

Thats some fine coding there lou.


What would you do differently?
 
2012-07-01 07:35:33 PM  

devioustrevor: I wonder if all these huge earthquakes over the past decade are going to effect stuff like this more often. I seem to recall reading that the earthquake that caused the tsunami devastating Japan last year made the year something like 4 seconds longer. The the Boxing Day tsunami from a few years back apparently had an even bigger impact.


No.
www.willoughby-eastlake.k12.oh.us

"Over the course of a year, the length of the day increases and decreases by about a millisecond, or about 550 times larger than the change caused by the Japanese earthquake.


A short article on it.
 
2012-07-01 09:04:51 PM  

MusicMakeMyHeadPound: What would you do differently?


Fix it.

Because NTP has been an establish protocol longer then I have been farking alive.
 
2012-07-01 11:11:20 PM  

fluffy2097: MusicMakeMyHeadPound: What would you do differently?

Fix it.

Because NTP has been an establish protocol longer then I have been farking alive.


That's rather circular. He did propose fixing it by removing the printks for the leap second on older RHEL kernels. You objected to that fix. So what would you have done differently?
 
2012-07-01 11:24:27 PM  

Splinshints: fluffy2097: If you are an admin that's always on-call, you've done it so very wrong.

There's a difference between being "on-call" and being "called on". I'm technically on call 24/7, but I haven't actually had to go in outside normal business hours for anything unplanned in a very long time. I think the last time was well over a year ago when I started getting weird temperature problem reports from a PBX and couldn't figure out why remotely.

/ turned out the probe was failing


back in the day, before I gave up the glamorous life of industrial automatons and installing backbones for a life of professional brewer, I had an apprentice that was wiz bang on all the server stuff, programing and all that.He was the son of our lead engineer and we gave the kid a summer job.
just to mess with the kid we crawled up into the over heads and nicked a cat wire and told him to find the problem.

took him five days if trying to get that server to talk before he actually bothered to check the wires.


this kid of course makes far more money than I will probably ever hope to see. but at least he knows to check for broken wires now :D

yeahyeah, CSB
 
2012-07-01 11:45:25 PM  

Cerebral Knievel: took him five days if trying to get that server to talk before he actually bothered to check the wires.


The fark? Seriously? There are status lights! There are even programs that show you virtual status lights!
 
2012-07-02 12:25:21 AM  

fluffy2097: Cerebral Knievel: took him five days if trying to get that server to talk before he actually bothered to check the wires.

The fark? Seriously? There are status lights! There are even programs that show you virtual status lights!


very bright kid.

just at the time, he had a problem with spotting the obvious. also, the wire wasn't severed, It was just cut enough to show that it was connected but it couldn't talk.
it was the kids dad's idea.

keep in mind, this was the mid 90's and the kid was 17, and never pulled a wire in his life before then.
 
2012-07-02 01:04:34 PM  

Cerebral Knievel: this kid of course makes far more money than I will probably ever hope to see. but at least he knows to check for broken wires now :D


When I took the exam for my CCNA class (a final exam at the school, separate from the cert exam) I had originally been scheduled to come in during a weekday afternoon. One of the other students I took the class with had run over on their time, and I got bumped to a Saturday afternoon. Still being in high school, I was naturally quite annoyed. I complained that I shouldn't have to give up my weekend because of their difficulties, etc.

So finally I go to take my exam. I spent FOUR HOURS trying to figure out I couldn't get two routers to talk to eachother over a ?parallel? (it's been awhile) cable... the cable lit up 'link good', but it had a bad wire somewhere that didn't allow communication. To this day, I still don't believe those little lights when there's an issue.
 
2012-07-02 03:08:20 PM  

rosonowski: Cerebral Knievel: this kid of course makes far more money than I will probably ever hope to see. but at least he knows to check for broken wires now :D

When I took the exam for my CCNA class (a final exam at the school, separate from the cert exam) I had originally been scheduled to come in during a weekday afternoon. One of the other students I took the class with had run over on their time, and I got bumped to a Saturday afternoon. Still being in high school, I was naturally quite annoyed. I complained that I shouldn't have to give up my weekend because of their difficulties, etc.

So finally I go to take my exam. I spent FOUR HOURS trying to figure out I couldn't get two routers to talk to eachother over a ?parallel? (it's been awhile) cable... the cable lit up 'link good', but it had a bad wire somewhere that didn't allow communication. To this day, I still don't believe those little lights when there's an issue.


I agree those serial cables' connectors were notoriously prone to failure and doesn't help that in a class room they get clobbered by novice users constantly.

/You got four hours for yours?
//We had 1 hour and it was pass or fail
 
2012-07-02 03:11:45 PM  

italie: Wow, if just a second can cause that type of carnage, imagine how bad that Y2K38 bug is going to be...


FTF'Nix
 
2012-07-03 02:04:43 AM  

rosonowski: Cerebral Knievel: this kid of course makes far more money than I will probably ever hope to see. but at least he knows to check for broken wires now :D

When I took the exam for my CCNA class (a final exam at the school, separate from the cert exam) I had originally been scheduled to come in during a weekday afternoon. One of the other students I took the class with had run over on their time, and I got bumped to a Saturday afternoon. Still being in high school, I was naturally quite annoyed. I complained that I shouldn't have to give up my weekend because of their difficulties, etc.

So finally I go to take my exam. I spent FOUR HOURS trying to figure out I couldn't get two routers to talk to eachother over a ?parallel? (it's been awhile) cable... the cable lit up 'link good', but it had a bad wire somewhere that didn't allow communication. To this day, I still don't believe those little lights when there's an issue.


When I took the final lab for CCNA I didn't bother to do it. (my team members were idiots)
 
Displayed 45 of 45 comments

View Voting Results: Smartest and Funniest


This thread is archived, and closed to new comments.

Continue Farking
Submit a Link »
On Twitter





In Other Media


  1. Links are submitted by members of the Fark community.

  2. When community members submit a link, they also write a custom headline for the story.

  3. Other Farkers comment on the links. This is the number of comments. Click here to read them.

  4. Click here to submit a link.

Report