NASA: DOS Glitch Nearly Killed Mars Rover

[iLikeMyiMac]

TANFORD, CALIF. -- A software glitch that paralyzed the Mars "Spirit" rover earlier this year was caused by an unanticipated characteristic of a DOS file system, a NASA scientist said Monday.
The flaw, since fixed, was only discovered after days of agonizingly slow tests complicated by the limited "windows" of communication allowed by the rotation of Mars, said Robert Denise, a member of the Flight Software Development Team at NASA's Jet Propulsion Laboratory.
On Jan. 21, the Spirit rover stopped communicating with the teams on Earth, beginning a cycle where the rover would reboot itself, over and over. After days of tests, the team finally discovered on Jan. 26 that the issue was tied to what was originally reported as corruption inside the rover's onboard flash memory.link

Remember when you put expensive objects in space always check to make sure it works before you launch it.

 

[Apple Hobo]

The best NASA could come up with was DOS?!

BTW, was this the rover that had some kind of PPC?

 

[musicpyrite]

Ahahah, if NASA was dumb enough to use DOS, then they almost deserved to loose the rover.

I hope that this incident is a warning and a good wake-up call.

 

[iLikeMyiMac]

There are probably a lot of fingers being pointed at the guy who put the software on the rover.
But when you need high reliability why would you put DOS on it.

 

[jhu]

There are probably a lot of fingers being pointed at the guy who put the software on the rover.
But when you need high reliability why would you put DOS on it.

fyi, dos is still used in many industrial applications because it's been well tested, it works, and its memory footprint is next to nothing. also, i believe the mars spirit actually uses vxworks instead of dos.

 

[iLikeMyiMac]

I guess you're right jhu. It was more of a RAM problem than an OS.

 

[jhu]

well, like you said, make sure it works before launching it. it's more of a qa issue.

 

[Daveman Deluxe]

There's lots of versions of DOS out there. I doubt this is MS-DOS that they're talking about.

 

[Makosuke]

The best NASA could come up with was DOS?!

Keep in mind also that DOS does not equal MS-DOS -- there are many OSes called "DOS" that are well-tested and heavily used in embedded systems where their next-to-no-footprint nature is ideal. Where I work we've run hydrogen powered cars on embedded systems running DR-DOS, for example, and it's worked fine for years.

Now, if they'd been using WindowsCE, then there'd be some monster stupidity involved.

 

[Apple Hobo]

Well, I guess it's not too bad if it has a good version of DOS.

 

[Les Kern]

There are probably a lot of fingers being pointed at the guy who put the software on the rover.
But when you need high reliability why would you put DOS on it.

But who's going to be the guy who goes to Mars to do a hard reboot of OSX after the 10.3.5 update ???
NASA does dos based because it's simple and incredibly reliable.

 

[EminenceGrise]

i believe the mars spirit actually uses vxworks instead of dos.

That makes sense - vxWorks will run on top of a DOS file system, there are several such machines where I work (it may support other filesystems as well, but I'm not sure). vxWorks is widely used in the embedded systems market and has been around forever. It's not bad for a real-time UNIX - though it does have some arcane bits from a user perspective, you don't usually run it as a desktop system anyway. Just the kind of thing you want for a Mars rover I'd say.

 

[iMeowbot]

Right, there is not actual DOS involved. The flash card that ran into problems was formatted with FAT. FAT only supports a limited number of files in its root directory, and the software controlling the rover failed to take this into account.

This is just another case of code monkeys passing themselves off as engineers. Nothing to see here, move along.

 

[mattthemutt]

We've been to Mars?

 

[Mr. Anderson]

There is a lot of testing done before launch. Like any code, there's always a bug somewhere....this one just happened to be a bit more serious than some others.

Give the rocket heads at JPL credit, their 50,000,000 mile tech support managed to fix it. Next time you send a probe into space for a multi year journey I think you'll be less likely to condemn the others who have made it there before you.

D

 

[iMeowbot]

It's nice that there was someone around who could fix it, but that never should have been an issue to start with. If that was a medical or safety application, there would be a list of every system constraint and something to check that it doesn't get hit. They were bailed out by a default error handler, but that's getting a fall broken by the padding under the rug because no one checked that the safety net wasn't made of moldy cheese.

The QA going into building this stuff is seriously lacking, that's been evident for a few years running now. It's too bad NASA isn't regulated by some of the other govt. agencies, they wouldn't let something like that happen again.

 

[crazzyeddie]

It's too bad NASA isn't regulated by some of the other govt. agencies, they wouldn't let something like that happen again.

You mean like the geniuses at the FBI and the CIA who can't even stop some terrorists on our own planet? Or the SEC, who let Enron put thousands of people out of their jobs and lose their retirement? Yeah, you're right, government agencies ARE the answer

 

[wdlove]

The really important news is that they fixed the problem with Mars Rover. It was a learning experieince for the scientist's at JPL. Hopefully they will put the knowledge to good use.

 

[Timelessblur]

before people knock dos and blame it. There is a reason that nasa choose to use Dos.

First it is very small. 2nd it is the most relieble and stable OS out there. Mind you it not user friendly at all and is fairly hard to use but it is still extermly stable. OSX may be stable but it does not hold a candle to how stable and reible Dos is.

 

[EminenceGrise]

before people knock dos and blame it. There is a reason that nasa choose to use Dos.

First it is very small. 2nd it is the most relieble and stable OS out there. Mind you it not user friendly at all and is fairly hard to use but it is still extermly stable. OSX may be stable but it does not hold a candle to how stable and reible Dos is.

Hahahahahahaha. Thanks, best laugh I've had yet today.

 

[jhu]

before people knock dos and blame it. There is a reason that nasa choose to use Dos.

First it is very small. 2nd it is the most relieble and stable OS out there. Mind you it not user friendly at all and is fairly hard to use but it is still extermly stable. OSX may be stable but it does not hold a candle to how stable and reible Dos is.

it depends on what you mean. osx is much more infinitely more stable if a poorly written program is executed whereas dos will crash or something similar. however, dos is able to do so well in an embedded environment because its memory footprint is very small and any applications that are run are extremely well behaved, or at least should be once they get to market.

 

[Counterfit]

Doesn't anyone here remember Pro-DOS? It ran on *gasp* APPLES!
It's already been said in here, but I guess it bears repeating: MS-DOS WASN'T THE ONLY KIND OF DOS!

 

[cslewis]

I think nuclear power plants use some sort of 'fail-proof' system, I wonder how practical it would be for a rover.

 

[iFaulder]

Ahahah, if NASA was dumb enough to use DOS, then they almost deserved to loose the rover.

I hope that this incident is a warning and a good wake-up call.

I couldn't agree with you more.

 

[Sol]

So which is it?

fyi, dos is still used in many industrial applications because it's been well tested, it works, and its memory footprint is next to nothing. also, i believe the mars spirit actually uses vxworks instead of dos.

So which is it? DOS or VXWorks that suffered the memory error on the rover? I followed the progress of the Rovers for a couple of months and this is the first time I learned that DOS was used in this mission.