• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

Good trick everyone must know !!

Xilikon

Xilikon

[H]ard|DCer of the Year 2008
2FA
Joined
Oct 12, 2004
Messages
15,022
I bet a lot of us will see a box hang with that line : "[12:13:27] Folding@home Core Shutdown: FINISHED_UNIT" with no further progress. Often, rebooting or killing the client will cause it to trash the WU, which is unfortunate. For a long time, I wondered how to fix that without trashing the client until now and I finally managed to kick it back in gear without destroying so here's how :

Linux :

-Bring the terminal window.
-Type "ps -e | grep -i fah" and note the PID associated to one or more struck FahCore_a1.exe processes.
-Type kill -9 pid (where you put the PID of the struck core)
-Look if it finally resume further. Normally, when you get rid of any struck core processes, it will upload the finished work.

Windows :

-Open the task manager.
-Look for any struck core then click it and click on "End process".
-Do that until there is no more cores running. It should resume fine.

What is happening when we see "[12:13:27] Folding@home Core Shutdown: FINISHED_UNIT" without further progress is because the client is always trying to stop the cores without success. The only way is to forcefully kill them and the client cannot do that. My small guide will allow you to do it so the client can continue.

Hope this is helpful to avoid losing a completed WU worth points in the 4 digits for nothing. I will try to see if Stanford can write a new client with the ability to forcefully kill any struck core (maybe with a 10 min timer and if by 10 mins, there are running cores, kill it).

 
Good stuff to know! ;)

I normally use this command to see just the F@H processes in Linux:

ps -e | grep -i fah

 
Thanks, edited the guide with the right command. I used ps alone and was lucky to see just 6 processes with the fahcore in it. It's possible we can see 20-30 and the correct command help uncover the correct core processes.

 
It seems like the only system I have this problem on is my main system in which case I just start up Gnome System Monitor (looks a lot like Windows Task Manager) and use that to kill the cores.
 
Thanks for the guide! Definitely will help me in the future, I've lost quite a few WU's this way.
 
Doh! I just had this problem yesterday and lost a WU on my Ubuntu boxen. Never again. Thanks Xilikon :)

 
I had two of these over the weekend! Thanks for the guide. Now if I can just remember it next time it happens. How bout a sticky?
 
Glad I wrote that since right now, another box got stalled... Must be the heat getting intolerable :( I cannot wait to move to the condo to get AC.

 
Xilikon said:
Windows :

-Open the task manager.
-Look for any struck core then click it and click on "End process".
-Do that until there is no more cores running. It should resume fine.
Click to expand...

i had to restart my pc so i close one native SMP client and the VMware, and after my PC reboot, i restarted the SMP client but i got an error, a classic error IMO


Code: 
Note: Please read the license agreement (fah.exe -license). Further
use of this software requires that you have read and accepted this agreement.

If you see this twice, MPI is working
If you see this twice, MPI is working
4 cores detected
No directory settings found in registry. Using current directory...


--- Opening Log file [June 26 05:09:47]


# SMP Client ##################################################################
###############################################################################

                       Folding@Home Client Version 5.91beta6

                          http://folding.stanford.edu

###############################################################################
###############################################################################

Launch directory: D:\Folding@Home\FoldSMP1
Executable: D:\Folding@Home\FoldSMP1\fah.exe
Arguments: -advmethods -forceasm -verbosity 9

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.

[05:09:47] - Ask before connecting: No
[05:09:47] - User name: leSLIe (Team 33)
[05:09:47] - User ID: 2A15A7363A356C29
[05:09:47] - Machine ID: 5
[05:09:47]
[05:09:47] Loaded queue successfully.
[05:09:47]
[05:09:47] - Autosending finished units...
[05:09:47] + Processing work unit
[05:09:47] Trying to send all finished work units
[05:09:47] Core required: FahCore_a1.exe
[05:09:47] + No unsent completed units remaining.
[05:09:47] - Autosend completed
[05:09:47] Core found.
[05:09:47] Working on Unit 05 [June 26 05:09:47]
[05:09:47] + Working ...
[05:09:47] - Calling 'mpiexec -channel auto -np 4 FahCore_a1.exe -dir work/ -suf
fix 05 -checkpoint 15 -forceasm -verbose -lifeline 1916 -version 591'

[05:09:47]
[05:09:47] *------------------------------*
[05:09:47] Folding@Home Gromacs SMP Core
[05:09:47] Version 1.74 (March 10, 2007)
[05:09:47]
[05:09:47] Preparing to commence simulation
[05:09:47] - Ensuring status. Please wait.
[05:09:57] - Starting from initial work packet
[05:09:57]
[05:09:57] Project: 2665 (Run 0, Clone 787, Gen 16)
[05:09:57]
[05:09:57] Assembly optimizations on if available.
[05:09:57] Entering M.D.
[05:10:22]  on if available.
[05:10:22] Entering M.D.
[05:10:31] Couldn't open Go file
[05:10:31]
[05:10:31] Folding@home Core Shutdown: FILE_IO_ERROR
[05:10:31] Finalizing output
Folding@home Core Shutdown: FILE_IO_ERROR
Finalizing output

when this happens to me i just close the SMP client, wait a couple of minutes and then restart the client, and then i works properly. But this time i did what u said

-Look for any struck core then click it and click on "End process".
-Do that until there is no more cores running. It should resume fine

and the SMP (which was at around 16%) starts over with a new WU (a different protein) from scratch, from cero. :p

I have to use this only if the client has finished working :D
 
please note that a file IO error is going to trash your WU... on win boxes... I had a batch of WU's that all did this not due to overclocks or anything just up and decided to be picky!

 
Yeah, this is why it's best to not touch the client and only kill the cores. Let the client pick the slack and I made the suggestion in the FCF forum then kasson acknwledged this and will take a look. Being a programmer myself, it's a very trivial modification (add a timer and if the client is waiting for over 10 mins without any feedback from the cores, do a forced kill).

 
Thanks Xili!!

I've had this before and I've usually just killed the client. I've never lost a WU to killing the client, but obviously this is much safer.

 
What should I do if it's stuck there, and no cores are running? All I see is fah6 & fahmon running, no cores?
 
Ahh, only needed to type it in with the correct capitalization, then it found it!
 
Bump for a worthy thread. Just used this now and managed to save the WU. Thanks Xilikon for the tip!

 
Hey many thanks Xilikon, like the member above I had just trashed a WU also. Never again thanks to your post. Thanks again :D

FOLD ON!
 
This is the first time I have seen this post.

Thanks for the bump!
 
You must log in or register to reply here.
Back
Top