Project: 2619 Upon finishing/sending work folder not cleaed

[macmath]

With the past two of these, they finish ok and send ok, but they cannot erase the work folder. Then a new wu cannot be started without throwing out the contents of the folding folder and starting again. Please consider this a report about Project 2619, not a help request.

Thank you.

[toTOW]

Can you post the related part of the Fahlog.txt please ?

[macmath]

Can you post the related part of the Fahlog.txt please ?

Here is the part of the log file saying that the work folder was not cleaned. (By the way, I only have 3072 MB of memory, not 4294965248 MB)

Code: Select all

[02:25:31] + Attempting to send results
[02:25:31] - Reading file work/wuresults_08.dat from core
[02:25:31]   (Read 17426417 bytes from disk)
[02:25:31] Connecting to http://171.64.65.56:8080/
[02:43:48] Posted data.
[02:43:48] Initial: 0000; - Uploaded at ~15 kB/s
[02:43:50] - Averaged speed for that direction ~15 kB/s
[02:43:50] + Results successfully sent
[02:43:50] Thank you for your contribution to Folding@Home.
[02:43:50] + Starting local stats count at 1
[02:48:01] - Warning: Could not delete all work unit files (8): Core file absent
[02:48:01] Trying to send all finished work units
[02:48:01] + No unsent completed units remaining.
[02:48:01] + Attempting to get work packet
[02:48:01] - Will indicate memory of 4294965248 MB
[02:48:01] - Detect CPU. Vendor: GenuineIntel, Family: 6, Model: 15, Stepping: 11
[02:48:01] - Connecting to assignment server
[02:48:01] Connecting to http://assign.stanford.edu:8080/
[02:48:01] ***** Got a SIGTERM signal (15)
[02:48:01] Killing all core threads

Folding@Home Client Shutdown.

and here is the part of the log file that showed how, after I tried to get another wu going, it would not start up again.

Code: Select all

[00:50:25] 
[00:50:25] *------------------------------*
[00:50:25] Folding@Home Gromacs SMP Core
[00:50:25] Version 1.91 (2007)
[00:50:25] 
[00:50:25] Preparing to commence simulation
[00:50:25] - Ensuring status. Please wait.
[00:50:25] Project: 0 (Run 0, Clone 0, Gen 0)
[00:50:25] 
[00:50:25] - Writing "": (overwrite)Could not open "" for writing!
[00:50:25] Error: Could not write local file.  Exiting.
[00:50:30] - Shutting down core
[00:50:42]   Exiting.
[00:50:47] - Shutting down core

This happened with the following:
Project: 2619 (Run 3, Clone 736, Gen 0)
Project: 2619 (Run 4, Clone 785, Gen 0)
It seems like I received
Project: 2619 (Run 3, Clone 861, Gen 0)
to work on but lost it somehow in the shuffle to try to get going again this last time. In case it is of interest, here are some excerpts from the start of the first one so you can see the flags and the version of the executables (which I believe to be the most recent for OS X):

Code: Select all

Launch directory: /Users/leibnitz/Library/FoldingFolder
Executable: /Users/leibnitz/Library/FoldingFolder/fah6
Arguments: -local -advmethods -forceasm -verbosity 9 -smp 

[04:27:42] - Ask before connecting: No
[04:27:42] - User name: macmath (Team 1971)
[04:27:42] - User ID: ****************
[04:27:42] - Machine ID: 1
[04:27:42] 
[04:27:42] Loaded queue successfully.
[04:27:42] 
[04:27:42] - Autosending finished units...
[04:27:42] + Processing work unit
[04:27:42] Trying to send all finished work units
[04:27:42] Core required: FahCore_a2.exe
[04:27:42] + No unsent completed units remaining.
[04:27:42] Core found.
[04:27:42] - Autosend completed
[04:27:42] Working on Unit 08 [April 3 04:27:42]
[04:27:42] + Working ...
[04:27:42] - Calling './mpiexec -np 4 -host 127.0.0.1 ./FahCore_a2.exe -dir work/ -suffix 08 -checkpoint 15 -forceasm -verbose -lifeline 564 -version 602'

[7im]

Unless I am mistaken, the unable to delete files message is a known SMP bug, which is only cosmetic. The client does overwrite the files the next time around.

Check the Known Bugs sticky post at the top of the SMP Win section of the forum.

[anandhanju]

Hello macmath, welcome to the forums!

As yoiu noticed, the client sometimes is known to detect the memory incorrectly and set it to a maximum possible value. IIRC, Project: 0 (Run 0, Clone 0, Gen 0) [not a real WU] is a manifestation of this. Reconfigure your client and specify a sensible amount of memory that you'd like to allocate for FAH (1024[M] perhaps?).

As 7im suggested, the message about unable to delete work files is observed to be a false warning. After you set the correct value for memory and give it a shot, and if you observe stray files in the work directory, post the file names/extensions here; maybe someone can figure it out.

[bruce]

If the real question is whether the WORK file is being cleared out or not, post a list of the contents of that directory along with the log showing the start and end of the WU and we'll all be able to figure it out.

[macmath]

If the real question is whether the WORK file is being cleared out or not, post a list of the contents of that directory along with the log showing the start and end of the WU and we'll all be able to figure it out.

Unfortunately, I've deleted the old work folders since. I thought it was safe to do so because the wu did complete, was accepted by the server, and the points were tallied for it. Beforehand I did look inside and it had 20-some files in it. If it happens again I'll keep better records. Right now one core of a mid-2007 iMac is at work on a 3044 and the other on a 3046, despite having used the flags
-local -advmethods -forceasm -verbosity 9 -smp
so it will be about a week before I get another shot at a 2619. [When I saw that I had received the 3044, and only one processor was busy, I requested a wu for the other processor.]

I apologize for not checking the stickies first to see what was already known about this specific issue. Thank you for your interest and help. I'll look into seeing how to assign a limited amount of memory to it. Is 1 GB a good size?

Thank you again.