Strange folding problem.

[tsuehpsyde]

Okay, I'm not sure exactly what's going on, but I'm at my whit's end. I borg absolutely every machine I can get my hands on that I'm allowed to, including the server I run SourceKills on. On my previous host I was able to fold fine, as it puts a nice value of 19 on the PID so it doesn't inturrupt anything with a normal priority or above. Since I moved to a new host with Debian, I cannot for the life of me get it to run. It's strange because I have a 24/7 folder in Florida running the exact same OS, and it folds all day every day on a box that's mostly idle (a server for a Linux User Group, 2.8GHz P4). Here's what the logs do EVERY time, and yes, I did chmod -R 777 folder/ to make sure everything was completely writeable/readable/executable.

[04:58:01] Core required: FahCore_82.exe
[04:58:01] Core found.
[04:58:02] Working on Unit 02 [December 18 04:58:02]
[04:58:02] + Working ...
[04:58:02]
[04:58:02] *------------------------------*
[04:58:02] Folding@Home PMD Core
[04:58:02] Version 1.03 (September 7, 2005)
[04:58:02]
[04:58:02] Preparing to commence simulation
[04:58:02] - Looking at optimizations...
[04:58:02] - Files status OK
[04:58:02] - Expanded 87938 -> 563399 (decompressed 640.6 percent)
[04:58:02]
[04:58:02] Project: 1850 (Run 5, Clone 177, Gen 12)
[04:58:02]
[04:58:02] Assembly optimizations on if available.
[04:58:02] Entering M.D.
[04:58:09] Protein: p1850_Myosin6_PT_US
[04:58:09]
[04:58:09] Completed 0 out of 500000 steps  (0%)
[05:13:08] Writing checkpoint files
[05:28:10] Writing checkpoint files
[05:43:21] Writing checkpoint files
[05:52:01] Writing local files
[05:52:01] Completed 5000 out of 500000 steps  (1%)

*snip*

[17:22:44] Completed 495000 out of 500000 steps  (99%)
[17:30:35] Writing checkpoint files
[17:45:45] Writing checkpoint files
[18:01:12] Writing checkpoint files
[18:14:55] Writing local files
[18:14:55] Completed 500000 out of 500000 steps  (100%)
[18:14:55] Writing checkpoint files
[18:15:55]
[18:15:55] Finished Work Unit:
[18:15:55] Leaving Run
[18:15:57] - Writing 414099 bytes of core data to disk...
[18:15:57]   ... Done.
[18:15:57] - Shutting down core
[18:15:57]
[18:15:57] Folding@home Core Shutdown: FINISHED_UNIT

And that's exactly where it stops. It just shuts down, no reason, no error, just done. I check my F@H stats and no additional processors are added so it never submits the WU, so I fire it back up to see what it does, and this is what it says.

[00:32:15] Loaded queue successfully.
[00:32:15] + Benchmarking ...
[00:32:21]
[00:32:21] + Processing work unit
[00:32:21] Core required: FahCore_82.exe
[00:32:21] Core found.
[00:32:21] Working on Unit 02 [December 22 00:32:21]
[00:32:21] + Working ...
[00:32:21]
[00:32:21] *------------------------------*
[00:32:21] Folding@Home PMD Core
[00:32:21] Version 1.03 (September 7, 2005)
[00:32:21]
[00:32:21] Preparing to commence simulation
[00:32:21] - Looking at optimizations...
[00:32:21] - Created dyn
[00:32:21] - Files status OK
[00:32:21]
[00:32:21] Folding@home Core Shutdown: MISSING_WORK_FILES

I had this happen once and I thought it may be a kernel problem, but I recompiled to an optimized gaming kernel and it still did it, and I tried making everything writable and I still get it. I've setup F@H on a few linux servers, and they all run the same way: in a screen in the background with mainly default values so they don't use up server resources like memory. For what it's worth, the server is an AMD AthlonXP 3000+ with a gig of memory.

This server might only put out a unit every couple of days, but they are idle cycles not being used that help the cause, and it's really frustrating to not be able to put those cycles to use, especially since I pay for them monthly. Idunno how many linux folders we have around here, but I'm hoping we got a few.


Full Folding Ahead!

 

[racercarl]

Are you using FAH502-Linux or FAH504-Linux on this box?

Have you deleted the cores and re-installed them?

are you running them as root or as a user?

do you have the proxy setting correct?

And if so, you have just passed my Linux skills

Unhappy Mage or Arkaine23 are the real pro's on this, If they don't pop on and answer try PM-ing them, they both have a mini distro for folding and have always been very willing to help out.

 

[unhappy_mage]

Can I get an 'ls -lR' of the FAH directory? You might also try using the finstall script, which simplifies installing FAH quite a bit.

 

[tsuehpsyde]

I'm using version 502, the same one I use on all of my folders. Yes I deleted the entire folder and tried again, same thing. Folding is running as a regular username. Proxy setings aren't needed.

As far as the directory.

tsuehpsyde@server1:~/f@h$ ls -lR
.:
total 6184
-rwxrwxrwx  1 tsuehpsyde tsuehpsyde  249236 Aug 20  2004 FAH502-Linux.exe
-rwxrwxrwx  1 tsuehpsyde tsuehpsyde   48057 Dec 21 19:32 FAHlog.txt
-rwxrwxrwx  1 tsuehpsyde tsuehpsyde 3440948 Dec 17 01:38 FahCore_79.exe
-rwxrwxrwx  1 tsuehpsyde tsuehpsyde 2542532 Dec 17 23:53 FahCore_82.exe
-rwxrwxrwx  1 tsuehpsyde tsuehpsyde    1489 Dec 17 01:38 MyFolding.html
-rwxrwxrwx  1 tsuehpsyde tsuehpsyde     105 Dec 17 01:38 client.cfg
-rwxrwxrwx  1 tsuehpsyde tsuehpsyde       8 Dec 17 01:38 machinedependent.dat
-rwxrwxrwx  1 tsuehpsyde tsuehpsyde    7168 Dec 17 23:53 queue.dat
-rwxrwxrwx  1 tsuehpsyde tsuehpsyde      46 Dec 17 01:38 start
-rwxrwxrwx  1 tsuehpsyde tsuehpsyde     157 Dec 21 13:15 unitinfo.txt
drwxrwxrwx  2 tsuehpsyde tsuehpsyde    4096 Dec 21 19:32 work

./work:
total 1156
-rwxrwxrwx  1 tsuehpsyde tsuehpsyde  49321 Dec 21 13:14 current.xyz
-rwxr-x---  1 tsuehpsyde tsuehpsyde    233 Dec 21 19:32 logfile_02.txt
-rwxrwxrwx  1 tsuehpsyde tsuehpsyde 693220 Dec 17 20:40 wuresults_01.dat
-rw-r--r--  1 tsuehpsyde tsuehpsyde 414099 Dec 21 13:15 wuresults_02.dat
tsuehpsyde@server1:~/f@h$

As far as the finstall script dunno what that is. I just have a start script in there that throws the FAH executable into a screen so I can screen into it to see how it's doing. Same setup I use on two other linux servers and they run w/o issue.

Full Folding Ahead!

 

[unhappy_mage]

Finstall does this all automatically. I was going to do it all by hand with my latest install, but then I decided to give this a try. It's really quite slick.

The permissions et al look good, that shouldn't be a problem. I assume FAH is running as "tsuehpsyde"?

 

[tsuehpsyde]

Yep, runs as tsuehpsyde. I'm completely stumped as to why it's consistantly failing. It just quits as soon as it finishes the work unit.

Full Folding Ahead!

 

[tsuehpsyde]

Okay, running with finstall, let's see if it runs w/o erroring this time. Quite the little script.

Full Folding Ahead!

 

[tsuehpsyde]

Hmm, I'm not sure what it did differently, but it appears to be folding fine now. Althought, FAHlog.txt is ALOT more verbose now. Either way, I'm glad she's workin, thanks mage.

Full Folding Ahead!