I've been getting a lot of these.

[SETI@home] Unrecoverable error for result
20ja05ab.24478.2529.786054.40_3
(process exited with code 2 (0x2))

Yes, Paul, I've checked your page.

[ADDENDUM] This is happening on the PowerPC 603ev 160MHz with 136MB RAM (+137MB swap) on a stripped-down Yellow Dog 3.0.1 Linux installation. Deatils are here. Source code was compiled with gcc 3.2.2 from various nightly zips (currently 2005-IV-03).

It has no GUI, no server, no graphics, no I18N, no D11N... the largest items I have on it are the gcc libs and emacs. df says I've got 44% of the disk available. When I /sbin/init 1 I have 24MB are free in RAM, 136MB in swap space (1MB used). And even when I had the KDE suite installed I still had enough swap free to run SETI, so I've ruled out the RAM and HD as culprits.

---

DO not know about the error message but you got trouble in river city, you have 4 computer listed on your computer list as follows

681626 1 28.19 3,249.91 Power Macintosh PowerBook3,5 Darwin 7.8.0
695576 2 0.00 0.00 Linux 2.4.22-2a
708973 3 0.00 0.00 Linux 2.4.22-2a
709020 4 0.00 0.00 Linux 2.4.22-2a
The first 2, 2.4.22-2a are not to be found, but the 3rd one has lots of wus, client error and noisy ones make up the bulk at a quick glance.....

---

Semper Eadem
So long Paul, it has been a hell of a ride.

Park your ego's, fire up the computers, Science YES, Credits No.

That 3rd ``Linux 2.4.22-2a`` is the machine in question - The rest have been merged into CPU ID 709020.

BTW: Good to see you, Pascal.

---

I've taken the machine offline. I can't tell if I ed WUs 845679 and 305108 or if they were already FUBAR when I got 'em.

I was hoping to make it a dedicated cruncher, but the way things are going, it's probably better I don't.

...well, at least until I can figure out what's is goin' on.

---

> ...well, at least until I can figure out what's is goin' on.
>
Those WUs are a couple losers, You'll notice they had problems from way back. However, your puter has 77 results attributed to it in the last 5 days of it's existence. You're getting NO credit(except for 3, and those claimed .04 .02 and .01 credits), but you are consistantly getting "Client error-Computing" errors on 97% of them.

Fixing this puter is definitely called for.

tony

---

> I've taken the machine offline. I can't tell if I ed WUs
> 845679 > and 305108
> or if they were already FUBAR when I got 'em.
>

These 2 wu is part of a bunch of wu that have been "stuck" for many months, and is at last passing through the system, but it's highly likely all of these old wu will error-out.

All the other wu your computer have crashed on the other hand is new wu, so you have a problem somewhere...

But what is a SETI error 2? Is it out of RAM? Is it looking for an X display? Does SETI say "I'm gonna take too long" and just gives up? Does it need a Prozac suppository?? I mean... <B>WTF?!?</B>

The machine hasn't crashed, hasn't thrown a single exception, the compile went fine and smooth, the logs show nothing out of order... It's done everything I've asked of it, including compiling and running other apps. Slowly, true, but like a good machine, it does what it's told.

---

> The machine hasn't crashed, hasn't thrown a single exception, the compile went
> fine and smooth, the logs show nothing out of order... It's done
> everything I've asked of it, including compiling and running other
> apps. Slowly, true, but like a good machine, it does what it's told.
>
It may do exactly what you ask it to, but it's certainly not doing what Seti is asking it too. I hope someone comes along that can help you with this Computation error. It's above my skill level.

I wish you the best

tony

Thanks, Tony.

---

> Thanks, Tony.
>
Your welcome,

the only thing I know about Linux is that he carried his blanket and sucked his thumb alot. (comic strip-Peanuts). Oh.....Wait....that's Linus.

Torvalds, not Van Pelt!

I think you should stop reading my posts for a while... they're affecting you too much, Pilot.

Now where's Chiana?

---

I need an answer, guys!

Forum full of geniuses... c'mon - Don't let me down!

Please?

---

Going out on a limb here, but I doubt it's a specific Seti error. More like the code at which the exitthread() integer is being called upon on whatever language the Seti crunching program is being written in. (C++ ??)

Yet what is behind error number 2, is as good a guess for you as for me. ;)

Your new error -108 (last two units) is a specific BOINC/Seti error, as shown here.

---

I know what the -108 is because the results page sates...

<file_xfer_error>
  <file_name>02ja04ab.18465.6785.554814.223</file_name>
  <error_code>-108</error_code>
  <error_message>MD5 computation error</error_message>
</file_xfer_error>

...but something like thisworkunit's response just says...

<message>process exited with code 2 (0x2)
</message>

I have no way of figuring out WTF that means without having to call for help.

Anyway, what makes me think that it's SETI-specific is that the error message in stdout contained the line [SETI@home], and not [---].

---

Ofcourse it's the Seti application that's killing the thread with an exit code of 2. But that's just because it's written do do so and showing that exit code for the programmers.

If it were Einstein@Home exiting with code 2, you'd a) post this on the E@H forums and b) find E@H exiting at code 2. Same code, same thread, different name. ;)

But it's not needed that it's a specific Seti error. Or BOINC error, for that. If it were happening in Windows, I'd say to check in DrWatson. So for Linux, check a specific debugger. Maybe that that can make heads or tails from it. ;)

---

it's not needed that it's a specific Seti error. Or BOINC error, for that.
Not to sound too sarcastic, but that's reeeeeeeeal helpful of 'ya. SETI is exiting with code 2, which may or may not be my problem, or may or may not be a bug.

Reeeeeeeeal helpful. I can now narrow it down to one of 100,000 possibilities... :-/

All I want to do is crunch. That's all. Maybe occasionally use the keyboard attached to the CPU for the occasional bout of programming (it's a more comfortable one than the laptop's). But I want that box crunching and returning valid WUs.

[sigh]

Is it really too much to ask for?

---

> it's not needed that it's a specific Seti error. Or BOINC error, for
> that.
> Not to sound too sarcastic, but that's reeeeeeeeal helpful of 'ya.
> SETI is exiting with code 2, which may or may not be my problem, or may or may
> not be a bug.
>
> Reeeeeeeeal helpful. I can now narrow it down to one of 100,000
> possibilities... :-/
>
> All I want to do is crunch. That's all. Maybe occasionally use the keyboard
> attached to the CPU for the occasional bout of programming (it's a more
> comfortable one than the laptop's). But I want that box crunching and
> returning valid WUs.
>
> [sigh]
>
> Is it really too much to ask for?

If we were talking about a Wintel environment, I'd expect this to be "file not found" but I don't know what errors would be thrown by GCC on a PowerPC.

... but I would expect to find that in the GCC documentation, probably.

---

Considering that it's SETI what defines the error codes --

.o0( ! ) [tap tap tap tappity tap]

I read through GCC's error.h and chased down all the #include<>s. They eventually all point to /usr/include/asm-ppc/errno.h. It states in the "enumeration" that #define ENOENT 2 /* No such file or directory */.

OK. What file or directory would it be looking for? Is there any way to that hunt down in the SETI sources?

---

> Is there any way to that hunt down in the SETI sources?
>
That's what I meant. Sorry for the confusion, but by saying it's not needed to be Seti specific, I meant it's not specific to the Seti cruncher program, but more like the original code and the language it was written in. So even though it is compiled to be used on Linux or on a Mac or in Windows, all those ors all point to different options; memory, swap file, a hitch in the unit.

The exitthread() in the original code specifies what will terminate the cruncher and throw out that code. So the original programmers can find why this workunit died on you or the cruncher died on this workunit, and enhance the program.

---

That's what I meant. Sorry for the confusion
It's alright - I've been known to do that too.

The exitthread() in the original code specifies what will terminate the cruncher and throw out that code.
Aha.

So the original programmers can find why this workunit died on you or the cruncher died on this workunit, and enhance the program...
...or repair the machine if necessary. And if I read errno.h right, it's looking for something that isn't there... [tap tap tappity tap tap] Oh...

[tap tap] .


I found the "error": A typo in the SETI app's name.

Let's see if it works now.

---