Y2038 bug may hit Unix, Linux machines
|
| Posted
:
Sat, 07 May 2005 01:24:00 GMT |
| By
:
Thomas Blythe |
| Category
:
Technology |
| News Alerts by
Email ( click
here ) |
|
|
|
|
|
CALIFORNIA: After the Millennium bug for which several billions of dollars were committed for research and updations in computer systems the world over, there is yet another bug on the horizon. It is the Year 2038 bug that is slated to hit computer users in that year.
To be precise, on Tuesday, January 19 03:14:07 2038, machines prone to this bug will alter calendars to go back to Friday, December 13 20:45:52 1901.
Computer programmers predict that this can result in incorrect and wildly inaccurate dates being reported by the operating system and applications. It is likely to cause serious problems on many platforms, especially Unix and Unix-like and Linux platforms, because these systems will "run out of time". They are reluctant to predict the extent of the damage.
What is special about this date? It is explained that Unix and similar operating systems do not calculate time based on the Gregorian calendar. Instead, they are known to simply count time in seconds from their arbitrary "birthday", that is, GMT 00:00:00, Thursday, January 1, 1970. The accepted practice among software programmers is to use a 32-bit variable for this number (32-bit signed time_t). The largest possible value for the end integer in this calculation is 2**31-1 = 2,147,483,647. So, 2,147,483,647 seconds after Unix's birthday falls on Tuesday, January 19, 2038. And one second later, theoretically Unix systems will revert to their birth date (like an odometer switching back from 999999 to 000000).
Experts are of the opinion that Linux users will be the hardest hit, because of the wider acceptance of this OS for its security and cost features. They are feared to grind to a virtual halt or go into a loop. This Linux's own Y2K nightmare can be more damaging than the Y2K bug, because the latter basically involved applications while the 2038 bug affects the time-keeping function itself.
Linux gurus are apprehensive about the bug's impact on the embedded field, where software does not get replaced frequently. As such, major telecom gadgets and equipment will be greatly affected. However, one ray of hope is that the 32-bit processing can be replaced thus overcoming the impact of the bug -- definitely before 2038.
But, the optimism must end there. The bug can have severe impact on records created today with calculations going beyond 2038, like insurance policies. There could be error messages splashing on Unix and Linux screens then. And Linux is getting to be the popular operating system these days.
Experts say one and sure-short way to overcome the problem is to switch over to 64-bit or longer time_t data storage. Some of the existing 32-bit codes can be changed and the programs recompiled. However, all these are not very easy tasks.
Copyright,
respective author or news agency
|
|
|
|
|
|
| |
|
Y2038 bug may hit Unix, Linux machines i don't think so.
By:
Bhavin Patel ,
Tue, 10 May 2005 05:14:14 GMT
|
|
Hello folks,
I want to point some of the things regarding this issue.
1. In 30 years there will be very few 32bit processer running in this world.
2. time_t data type is component of c language. So software written using c language can face such problem on windows or any other os which have 32bit integer.
My question to author is why this problem is shown only context to *nix problem not c language problem or limitation of 32bit date time datatype problem?
Bhavin
|
|
I'm confused...
By:
Raymond Hilliard ,
Mon, 09 May 2005 15:26:12 GMT
|
|
Can someone please tell me what the difference between Linux/Unix systems and my Windows XP?
I got the news that Internet Explorer had security issues so I switched to the Mozilla Firefox browser. Am I really any better off?
I am new(obviously)to all this.
And I do not understand why anyone has a problem with the 2038 bug since 32-bit technology is already out there.
I am really happy I took the time to read all these "Have Your Say" items. It is a learning experience!
Confused (but not ignorant), Raymond
|
|
If only this problem were MORE widespread...
By:
Sgt_Jake ,
Mon, 09 May 2005 14:24:28 GMT
|
|
Listen... If my computer is still running as an un-re-booted, un-recompiled and un-patched system 33 years from now, I'm one [REALLY] happy customer.
If only all operating systems were at a POTENTIAL risk of crashing during that time frame.
|
|
LOL
By:
Dumb & Dumber ,
Mon, 09 May 2005 11:11:36 GMT
|
|
Oh not...what a sheer idiocity again...
According to my knowledge nowadays almost all unix/linux versions are using 64bit to count time. And of course somebody will use these softwares and architectures by 2038.
Please have a look man! Could you see any advancing in the technology?
|
|
foolish article
By:
Skunk Worx ,
Mon, 09 May 2005 01:29:48 GMT
|
|
i was building/wirewrapping and testing 8 bit computers in 1983. that's 22 years ago.
then 16 bit, then 32 bit in rapid succession...and now several people in my neighborhood have 64 bit machines running 64 bit linux.
oracle, db2, postgres, mysql...all the linux databases that matter will have all their time fields properly migrated into 64 bit within the next few years.
in the mean time, countless microsoft users will have their browsers hijacked and their credit card numbers exposed and used to buy blow-up sex dolls and fleshlights in the (former) soviet russia.
what a terrible, misinformed, ignorant troll article.
|
|
scaremongering
By:
Linuxadmin ,
Sun, 08 May 2005 23:48:29 GMT
|
|
Most widely-used Linux distributions already have a version that's been compiled for 64-bit systems and is therefore invulnerable to this bug. Anyone who buys an AMD64 processor and downloads a freely available Linux distro is invulnerable to this bug already.
Its ridiculous to think that in more than 30 years (a million years in the computer industry) that people will still be using computers which have this problem when many computers sold to consumers today are not vulnerable to it.
It also disproves the authors point that "all these are not very easy tasks", they've already been done for thousands of applications (modern Linux distros can have up to 10,000 software packages in their repositories) and often not by the people who originally wrote those applications (most packagers are 3rd parties to the authors).
Another point: "The bug can have severe impact on records created today with calculations going beyond 2038, like insurance policies". Total crap - if this was happening the results could be seen straight away when you get the paperwork back from your bank stating that your repayments will be due until '9 may 1910' or something. Almost all bank or financial systems will be built with safeguards such as using very long variables and possibly stuff like arbitrary precision maths libraries - the risk of losing money is just too great for them not be very careful. Even the humble spreadsheet program on Linux (OpenOffice.org Calc) handles dates beyond 2038 fine, even if the underlying operating system doesn't.
|
|
Y2K38 bug
By:
Sid Boyce ,
Sun, 08 May 2005 20:06:19 GMT
|
|
Chicken Little promised us Y2K doom, now we hear the same squawk. I enjoyed Y2K, my company paid me an extra £1000.00 for doing nothing.
Unfortunately, I'll be dead or too old by 2038, but trust me Chicken Little, it's all a filthy rumour and your great grandchildren will me ribbing you or someone else's grandpa something rotten about you running a 32-bit system.
Al least you have one follower so far, may your new religion grow, we have long needed a successor for the International Flat Earth Society.
|
|
Linux lovers hate critics
By:
DaleW ,
Sun, 08 May 2005 16:34:48 GMT
|
|
Linux lovers leap out from under the bed and off the wardrobe as soon as a problem, flaw or potential error is mentioned for their beloved OS.
If it was Windows with a problem, whether now or one reported for 100 years in the future every linux nut gets in line to knock them down.
As a PC OS Windows is better - face facts
For a server, I would go with Linux
Get a grip gnumber9, it's a news story, some people are interested and can have an opinion even if it does say something bad about linux, heaven forbid..
|
|
old news... troll like article
By:
gnumber9 ,
Sun, 08 May 2005 16:16:16 GMT
|
|
The author is either trolling for a kneejerk or is a complete idiot or both.
Making the user MORE aware of the "year 2038 bug" isn't going to do any good, since most users are not programmers! Programmers are aware and have a 30 year deadline.
Of course the asteroids are going to rain on us and we'll be in a new ice age due to global warming and WWW IIII will all occur before this date has been reached. Just read this web site's news.
|
|
This is foolishness!
By:
Will Morrison ,
Sun, 08 May 2005 14:27:19 GMT
|
|
Oh, come on! First it was social secuurity going under in 45 years, so that is an imminent crisis, now it's this. What C***! In 33 YEARS we will have an issue? Is there ANYONE out there that still wants to be running the same machine 33 YEARS from now?
Let's see, 15 years ago, not even HALF of the time in question, I was using a Commodore 64... Anyone out there still use one of those? Doing serious, mission critical things with it?
15 years ago, 8 bit computing was really great, why would anyone EVER need any more than 640k of RAM??? In 33 years, I would expect that things will be so far beyond this issue, that this is nothing more than Chicken little c***. 64 bit? Why wait there? It's taken far fewer years for us to go to 32 bits as a standard than what this article is talking about, and I see no reason for it to slow down to a dead halt now, which is what would be required for this to be a real issue.
This is just more FUD from guess who(?)!!!
|
|
Does it really matter?
By:
Anonymous ,
Sun, 08 May 2005 12:02:38 GMT
|
|
33yrs from now only if you're living somewhere behind gods back will this continue to be a problem.
|
|
Nicholas Donovan
By:
CEO ,
Sun, 08 May 2005 07:33:32 GMT
|
|
Regarding the issues of the year 2038 and computer systems going crazy reporting the wrong dates etc., let me assure you that this is simply not going to be the case.
All 32bit systems, be they Unix, Windows, Mac etc. are prone to this issue. However unlike proprietary systems, Open Systems can be solved with a simple recompile of the kernel_time date value as a 64 bit integer.
This is all mute of course as by the year 2038, 64 bit (and who knows, maybe even 128 bit architecture) will be the rule of the day so this whole scare is really rather pointless.
The fact that no author could be easily attributed to this article is also interesting.
Cheers,
Nicholas Donovan
-- --
Nicholas Donovan - CEO
Ioni Systems
Dallas, Texas USA
|
|
2038 Bug Only Affects 32-Bit Chips
By:
Kevin Sutherland ,
Sun, 08 May 2005 07:32:13 GMT
|
|
From what I understand, the 2038 bug is only expected to affect 32-bit chips and therefore, by the time the year 2038 comes around, all computers should well beyond 32-bit and onto 64-bit or even better. This problem has been known at least since the start of the new millenium is nothing new. *nix (e.g. Linux and Unix) machines are also no more prone to it then any other operating system.
More information on this problem can be found at:
http://en.wikipedia.org/wiki/2038_bug
|
|
2038
By:
lnx4u2 ,
Sun, 08 May 2005 06:35:32 GMT
|
|
poppycock i say were not scared bring it on!!!!!!!
|
|
|
|
|
