Linux box can't fold anymore...

H

HighYield

Gawd
2FA
Joined
Feb 2, 2008
Messages
944
My Ubutu Linux boxen has been a reliable folder. It is an E6700@3.1 with 2gig ram, It ran for a couple of months with Ubuntu 7.10 and then I upgraded (not a clean install) to 8.04. Since then it has run ok for several WU but recently it can no longer complete a unit. It is like it doesn't have permission to write to its own folder anymore. It completed the WU and then tried to shutdown.....and then hung. It never uploaded the work. I tried closing the program and restarting and now it won't run at all. I reloaded the software onto the machine in a new folder with a new folder name and it ran for another 4 WUs and now this install is doing the same thing....hangs just after completion without sending the work back. I have seen similar behavior on Vista 64 with NOD32 virus protection enabled...but I don't run any virus protection with linux so this can't be the cause. I am half wondering if I shouldn't just wipe the hard drive and reinstall everything clean. I am Linux impaired so any suggestions how to return to reliable production would be appreciated.

From the log
...
[05:42:06]
[05:42:06] Project: 2605 (Run 14, Clone 62, Gen 50)
[05:42:06]
[05:42:06] Assembly optimizations on if available.
[05:42:06] Entering M.D.
[05:42:13] Rejecting checkpoint
[05:42:13] OPC
[05:42:13] Writing local files
[05:42:13]
[05:42:13] Writing local files
[05:42:14] Extra SSE boost OK.
[05:42:14] Writing local files
[05:42:14] Completed 0 out of 500000 steps (0 percent)
[05:53:47] Writing local files
[05:53:47] Completed 5000 out of 500000 steps (1 percent)
.......Snip....
[01:00:47] Completed 500000 out of 500000 steps (100 percent)
[01:00:47] Writing final coordinates.
[01:00:47] Past main M.D. loop
[01:00:47] Will end MPI now
[01:01:47]
[01:01:47] Finished Work Unit:
[01:01:47] - Reading up to 3723552 from "work/wudata_05.arc": Read 3723552
[01:01:48] - Reading up to 1777580 from "work/wudata_05.xtc": Read 1777580
[01:01:48] goefile size: 0
[01:01:48] logfile size: 16913
[01:01:48] Leaving Run
[01:01:52] - Writing 5522445 bytes of core data to disk...
[01:01:52] ... Done.
[01:01:53] - Shutting down core
[01:01:53]
[01:01:53] Folding@home Core Shutdown: FINISHED_UNIT ----HANGS HERE
Folding@Home Client Shutdown. ---Control-c
......restart FAH
Folding@Home Client Version 6.02beta
http://folding.stanford.edu
###############################################################################
###############################################################################
Launch directory: /home/john/Desktop/fah6b1linux
Executable: ./fah6
Arguments: -smp -forceasm
Warning:
By using the -forceasm flag, you are overriding
safeguards in the program. If you did not intend to
do this, please restart the program without -forceasm.
If work units are not completing fully (and particularly
if your machine is overclocked), then please discontinue
use of the flag.
[10:38:28] - Ask before connecting: No
[10:38:28] - User name: [H]ardApe (Team 33)
[10:38:28] - User ID: 22887B0A7C6C7BD5
[10:38:28] - Machine ID: 1
[10:38:28]
[10:38:28] Loaded queue successfully.
[10:38:28]
[10:38:28] + Processing work unit
[10:38:28] Core required: FahCore_a1.exe
[10:38:28] Core found.
[10:38:28] Working on Unit 05 [May 13 10:38:28]
[10:38:28] + Working ...
[10:38:28]
[10:38:28] *------------------------------*
[10:38:28] Folding@Home Gromacs SMP Core
[10:38:28] Version 1.74 (November 27, 2006)
[10:38:28]
[10:38:28] Preparing to commence simulation
[10:38:28] - Ensuring status. Please wait.
[10:38:28] Exiting.
[10:38:45] - Shutting down core

 
Perhaps a stray FahCore_a1.exe or mpi process still running? Tried a reboot?
 
This is one reason I never do upgrades. I originally did an upgrade from openSUSE 10.2 to 10.3 and it didn't work out too well.

Unless you have some hardware going bad, I wouldn't be surprised if the problem is with the upgrade.

 
SmokeRngs said:
This is one reason I never do upgrades. I originally did an upgrade from openSUSE 10.2 to 10.3 and it didn't work out too well.

Unless you have some hardware going bad, I wouldn't be surprised if the problem is with the upgrade.

Click to expand...

+1 x 10,000.

I have never, ever, never ever ever had good luck with an upgrade. Things get broken, and that is all there is to it.

Wipe the drive and try a fresh install of 8.04...My Ubuntu boxen have been crunching away just fine since a clean install a couple weeks ago...

Good luck man, I hope that solves your problems...

 
I have had good success with upgrading Ubuntu .. maybe try deleting the contents of your folding folder, redownload the client and give it a whirl .. ?


[F]old|[H]ard
 
ThreeDee said:
I have had good success with upgrading Ubuntu .. maybe try deleting the contents of your folding folder, redownload the client and give it a whirl .. ?


[F]old|[H]ard
Click to expand...

Yes, try that. Sometime, upgrades bork things. This happened 3 or 4 times myself with the latest occurence 2 weeks ago (one box chugging for 6 months fine then out of the blue, borks itself. Deleted the client folder, reinstalled and all is find and dandy).

 
Xilikon said:
Yes, try that. Sometime, upgrades bork things. This happened 3 or 4 times myself with the latest occurence 2 weeks ago (one box chugging for 6 months fine then out of the blue, borks itself. Deleted the client folder, reinstalled and all is find and dandy).

Click to expand...

I did try this as indicated in the OP and it worked for about 4 WU and then died again.

This problem was too much of a pain in the ass to troubleshoot. Seems like the consensus was too wipe the drive and do a fresh install of everything. Did that and then it still wouldn't run :eek:
I forgot you need to type this the first time:
sudo aptitude install ia32-libs
Now it seems to be working just fine. After it runs through a dozen WUs I'll start trusting it again.

 
I've actually had this occur several times in a couple of my Ubuntu 8.04 VM's. that's one of the reasons I check them at least twice a day. No particular reason that I've found, but it just seems like they stall once they upload the finished WU. Like you, I just do a control-C, wait a few seconds (about 10), and then restart. It goes well from there, and might not stall like that for several days.:confused: No idea why, though.

Edit - I always install the 32bit libs prior to loading SMP (sudo apt-get install ia32-libs).

 
If a box is having repeated issues like that, I guess the overclocking settings is too borderline to stability. It's a good idea to lower some or bump some voltage.

 
Xilikon said:
If a box is having repeated issues like that, I guess the overclocking settings is too borderline to stability. It's a good idea to lower some or bump some voltage.

Click to expand...

I might have thought that as well, but I'm running two VM's on each box, and only one of the VM's has the issue sometimes. It's never both of them, so I guess that's why I was thinking it might be an innue with the install or WU that it was processing. This occurs on a couple different boxes, too.

 
If that's a VM only, I guess it's a good idea to delete the whole VM instance and create a new one.

 
You must log in or register to reply here.
Back
Top