One in a million

[djp]

129 of 130 identically configured boxen are folding offline while they patiently await the release of the Phase II version of dfQ. Fortunately, they're merely P-3 1GHz w/256MB RAM boxes running the text console minimized under Windows 2000 (not a service, just -if and -rt), so the luckiest of them only has about 270 generations buffered so far.

One little box is giving me fits, though. It had folded for a while, but now whenever I launch FOLDIT.BAT , I see this on my screen:

Error Number  -490, 90910
Unrecognized Value
expr4parseValue

When I hit the Enter key, I get more info:

[NULL_Caption] FATAL ERROR:  [010.000]  Unable to continue
Hit Return

and then the program terminates.
This didn't tell me much, so I looked at the error.log file and saw this several times:

========================[ Jun 24, 2003  3:35 PM ]========================
ERROR: [010.000] {trajtools.c, line 2022} Could not open/create data file C:\DOCUME~1\admin\LOCALS~1\Temp\2_0792
FATAL ERROR: [010.000] {trajtools.c, line 2024} Unable to continue

only with the name of the temp file different each time.

Here's an abbreviation of my filelist.txt file:

fold_0_4s8cy36o_3825_protein.log.bz2
4s8cy36o_0_protein_0003828.val
.\fold_0_4s8cy36o_40_4s8cy36o_protein_1.log.bz2
.\4s8cy36o_0_4s8cy36o_protein_1_0000042.val

<snipped-out a bunch of filename pairs>

.\fold_0_4s8cy36o_10_4s8cy36o_protein_25.log.bz2
.\4s8cy36o_0_4s8cy36o_protein_25_0000012.val
CurrentStruc 0 1 126 26 0 0 10000000.000 10000000.000 -10000000.000 0.000 0.000 1.250 2.300 764.756 ------------------HHHHHHHH-HHHHHH----------------------HHHHHHH----------------HHHHH-------------
f2ec55182807cfa875b2030c73d9cc69

With an error telling me that I can't open/create a data file, my first assumptions would be that I'm either out of disk space or I lack permissions to create a file in this location. Neither possibility tests true, though. I've got 1GB free on the C: drive (where the error message says it wants to create a file) and 15.8GB free on the D: drive (where my distribfold directory lives). My distribfold directory contains 26 *.log.bz2 files, and there's been no change in security policies, so I must have been folding for a while successfully. (Kinda' rules-out the whole permissions thing, doesn't it?)

I haven't tried sneaker-netting the whole folder over to another PC and running it there yet. I'll try that in a few hours and either edit this message or post a follow-up.

I haven't yet tried hooking it up to the Internet and running with -ut to dump the work done so far. I'll try that, too.

Other tactics I have tried and failed with:
- Copied a bunch of garbage files onto the C: drive so that it would have less than 1.0 GB of free space in case the system was only checking the low-order bits of the freespace and getting fooled by EXACTLY a GB free.
- Overwrote all files with original installation in case of file corruption.
- Rebooted the system because.... Hey! It's MS-Windows. It likes to be rebooted.


I have renamed my distribfold folder and installed a new instance of the folding client and it works just fine for the moment. (I'll follow-up if it errors-out again.)

Does anyone have another suggestion for me?

[Brian the Fist]

I know where that error is coming from. Unfortunately I cannot look it up for a couple weeks as I dont have access to the book. In the mean time, try running it as a different user and see if that works. There is some problem with writing to the temp directory or the path of the temp directory it seems.

[Brian the Fist]

Is there any chance you still have this directory preserved so I could try it out on my own machine? If not Ill do my best without it...

The most likely cause is a buffer overflow happening somewhere in some string variable. Again, if you have the directory saved somewhere that'd be super helpful

[djp]

I'm sorry. I've kinda' let the quantity of DF work over-rule my concern for quality. I've just been trashing bad results and starting over again. To your credit, I haven't had that much to trash, comparitively speaking.

I haven't approached this project with the careful discipline I apply to my job in software QA. I can change that. I'll start today capturing anomaous behaviors and collecting them in my bug notebook.