Boinc gone mad

Message boards : Questions and problems : Boinc gone mad
Message board moderation

To post messages, you must log in.

AuthorMessage
PeteBroad

Send message
Joined: 23 Oct 13
Posts: 17
United Kingdom
Message 50980 - Posted: 23 Oct 2013, 15:59:41 UTC

O.K, anybody seen anything like this. Overnight Boinc has stopped running on one of my machines. I restarted the machine, same problem, the icon on the taskbar indicated it was connecting to the client but it never did. Did some checks and found to my amazement that from going awry at about 4 in the morning until 10 in the morning it had created some 12 million plus slots, What the hell. Goodness knows whats happened and deleting all the slots is going take for ever.


Pete
ID: 50980 · Report as offensive
Profile Jord
Volunteer tester
Help desk expert
Avatar

Send message
Joined: 29 Aug 05
Posts: 15477
Netherlands
Message 50981 - Posted: 23 Oct 2013, 16:05:14 UTC - in response to Message 50980.  

No, never seen that before. But even then, as ever, without information it can be anyone's guess.
ID: 50981 · Report as offensive
PeteBroad

Send message
Joined: 23 Oct 13
Posts: 17
United Kingdom
Message 50982 - Posted: 23 Oct 2013, 16:19:23 UTC

It's the latest version of boinc, updated about a month ago running on WinXP SP2, Intel dual core, Nvidia 460GT, 2gb Ram running various tasks. Machine first had boinc installed in December 2006 and has run without fault since then. What other info do you need?

Pete
ID: 50982 · Report as offensive
Profile Jord
Volunteer tester
Help desk expert
Avatar

Send message
Joined: 29 Aug 05
Posts: 15477
Netherlands
Message 50983 - Posted: 23 Oct 2013, 16:50:27 UTC - in response to Message 50982.  
Last modified: 23 Oct 2013, 16:51:45 UTC

'The latest version of BOINC' can still mean anything. Please give the version number. If you do not dare start BOINC, you can find this information by going:
C:\Program Files\BOINC, right-click on the boinc.exe program, Properties, Details tab. It'll state what the number is in the Product Version.

For you 'the latest version of BOINC' could mean 7.0.28 or 7.0.64, for the general Linux user it's either 7.0.27 or 7.0.65, for the testers it is 7.2.23 and then there's still a lot of versions in between. Depends on where you got it from as well.

Furthermore, it would be nice to know which project caused it. The slots directories are made by BOINC to write the project's science application and task information to, before they're executed in there. So if you still have one of your several million slots directories, please open a couple up and see if it's all one project or multiple ones. And tell which ones. Or tell which application's executables you see and we'll be able to derive from there what it could be.

Can you otherwise check stdoutdae.txt in your BOINC Data directory, find if it wrote at 4am what it was doing and post that here?

Do you have anti-virus? If so, have you made sure to exclude the BOINC Data directory from being actively scanned by that AV? Same question for other Anti-Malware programs.

Why XP SP2? Why not updated to SP3?
ID: 50983 · Report as offensive
PeteBroad

Send message
Joined: 23 Oct 13
Posts: 17
United Kingdom
Message 50986 - Posted: 23 Oct 2013, 19:22:28 UTC

Version 7.0.64 which came from here. The 12.4 million slots have now finally been deleted, took 9 hours, and I took a chance and restarted Boinc. I obviously lost any work in progress but on restarting it re-initialized all my projects (apart from a couple of dead ones) and picked up work. Unfortunately, I deleted a few other files including stdoutdae.txt in fact everything after the 04:15 entry for Volpex, which just looked to be an attempt to get work, the next entry was at 09:30 when I restarted Boinc.

The machine in question is part of an 8 machine farm and they all run pretty well the same projects, none of the other machines running a mix of Windows 7/8/XP SP2 and SP3 have been affected.

Pete
ID: 50986 · Report as offensive
Profile Jord
Volunteer tester
Help desk expert
Avatar

Send message
Joined: 29 Aug 05
Posts: 15477
Netherlands
Message 50988 - Posted: 23 Oct 2013, 20:25:35 UTC - in response to Message 50986.  

OK, I forwarded your plight to the developers. Perhaps that they can put a hard value on amount of slots to stop this behaviour.

May I still know which projects you run?
Or else, can you post the first 40 lines from your BOINC start-up messages? That way we can see those projects as well and perhaps check at the projects if there's any reason why BOINC went into this frenzy.
ID: 50988 · Report as offensive
Profile Jord
Volunteer tester
Help desk expert
Avatar

Send message
Joined: 29 Aug 05
Posts: 15477
Netherlands
Message 50997 - Posted: 24 Oct 2013, 9:10:33 UTC - in response to Message 50986.  

The developers ask that you add the following cc_config.xml file to your data directory (default at C:/Documents and Settings/All Users/Application Data/BOINC/ on XP and default hidden, so either set Windows Explorer to show hidden files and folders or put the path to it directly into Windows Explorer and hit Enter), and leave it there for now. Perhaps that the same problem never happens again, but if it does, at least you have some logging that shows what could have caused it. That's to say, if you then do not delete stdoutdae.txt or stdoutdae.log, of course.

<cc_config>
<log_flags>
<task_debug>1</task_debug>
</log_flags>
<options>
</options>
</cc_config>


If you already have a cc_config.xml file, add the above into it.
When there's no cc_config.xml (there's none by default, so don't worrry if you do not have one), you can make it by opening Notepad, putting the above lines into it, then do:
Click File->Save As...
-> Navigate to C:/Documents and Settings/All Users/Application Data/BOINC/
-> Set "Save As type" to All files
-> Encoding: ANSI
-> File name: cc_config.xml and then click Save.

Do check that it all went correctly, and that the file didn't get the .txt extension. If it did, rename it so it only has an .xml extension.

Next open BOINC Manager->(View->Advanced view->)Advanced->Read config files.
When you now open event log (CTRL+SHIFT+E), a lot more data will be written to it when tasks run. Including when they're started, in which slot, and what else happens to them. [/code]
ID: 50997 · Report as offensive

Message boards : Questions and problems : Boinc gone mad

Copyright © 2024 University of California.
Permission is granted to copy, distribute and/or modify this document under the terms of the GNU Free Documentation License, Version 1.2 or any later version published by the Free Software Foundation.