Thread 'Computing error'

Message boards : BOINC client : Computing error
Message board moderation

To post messages, you must log in.

AuthorMessage
Merciadri Luca

Send message
Joined: 1 Aug 09
Posts: 33
Belgium
Message 27880 - Posted: 11 Oct 2009, 13:19:45 UTC

Hello,

When I sometimes face OS freezes, or that I encounter electrical problems, my computer needs to be rebooted, just like any other one.

I am always surprised, after its reboot, to see at least one of the BOINC tasks which my CPU was working for to be completely cancelled, even after hours of previous ``correct computing.'' The problem is that if you encounter such a freeze, there is no reason to cancel the *whole* task: it is possible that your computer failed during the last 10 minutes, on maybe 50 hours of processing.

Thus, wouldn't it be nice to implement an automatic backup of this? If not so, why? I thought there was already such a mechanism, but it appears not to be here, in my version...
See http://www.student.montefiore.ulg.ac.be/~merciadri/ for more pieces of info.
ID: 27880 · Report as offensive
ProfileJord
Volunteer tester
Help desk expert
Avatar

Send message
Joined: 29 Aug 05
Posts: 15585
Netherlands
Message 27882 - Posted: 11 Oct 2009, 15:21:45 UTC - in response to Message 27880.  

Which BOINC version is this with?
On which operating system?
With which project or projects?
ID: 27882 · Report as offensive
Merciadri Luca

Send message
Joined: 1 Aug 09
Posts: 33
Belgium
Message 27883 - Posted: 11 Oct 2009, 16:31:51 UTC - in response to Message 27882.  

Which BOINC version is this with?

6.6.36
On which operating system?

Microsoft Windows XP Pro.
With which project or projects?

Projects: QMC@Home, SETI@home (not astropulse).

Note that I had the same problem on my Debian machines, with the same version.

See http://www.student.montefiore.ulg.ac.be/~merciadri/ for more pieces of info.
ID: 27883 · Report as offensive
ProfileJord
Volunteer tester
Help desk expert
Avatar

Send message
Joined: 29 Aug 05
Posts: 15585
Netherlands
Message 27884 - Posted: 11 Oct 2009, 17:44:00 UTC - in response to Message 27883.  

What kind of computation errors do you get? Can you post a link to one of the tasks with an error, or a link to the affected computer?

Are there any messages in the Event Log in Windows?
How did you install BOINC, as a normal installation or using the protected application execution setup (aka the Windows service)?
ID: 27884 · Report as offensive
Merciadri Luca

Send message
Joined: 1 Aug 09
Posts: 33
Belgium
Message 27908 - Posted: 12 Oct 2009, 11:37:44 UTC - in response to Message 27884.  

What kind of computation errors do you get? Can you post a link to one of the tasks with an error, or a link to the affected computer?

Unfortunately, the only one I have is: http://qah.uni-muenster.de/result.php?resultid=12027091.

Are there any messages in the Event Log in Windows?

No.

How did you install BOINC, as a normal installation or using the protected application execution setup (aka the Windows service)?


I installed it normally, but do you speak about the eXecute Disable Bit function?

Thanks.

See http://www.student.montefiore.ulg.ac.be/~merciadri/ for more pieces of info.
ID: 27908 · Report as offensive
ProfileJord
Volunteer tester
Help desk expert
Avatar

Send message
Joined: 29 Aug 05
Posts: 15585
Netherlands
Message 27911 - Posted: 12 Oct 2009, 12:16:43 UTC - in response to Message 27908.  

Unfortunately, the only one I have is: http://qah.uni-muenster.de/result.php?resultid=12027091.

This error is an application error. You're better of reporting it at the affected project forums.
ID: 27911 · Report as offensive
Merciadri Luca

Send message
Joined: 1 Aug 09
Posts: 33
Belgium
Message 27936 - Posted: 12 Oct 2009, 19:21:08 UTC - in response to Message 27911.  

Unfortunately, the only one I have is: http://qah.uni-muenster.de/result.php?resultid=12027091.

This error is an application error. You're better of reporting it at the affected project forums.

Thanks.

But is there actually an automatic backup for WUs?
See http://www.student.montefiore.ulg.ac.be/~merciadri/ for more pieces of info.
ID: 27936 · Report as offensive
ProfileJord
Volunteer tester
Help desk expert
Avatar

Send message
Joined: 29 Aug 05
Posts: 15585
Netherlands
Message 27940 - Posted: 12 Oct 2009, 20:22:41 UTC - in response to Message 27936.  

But is there actually an automatic backup for WUs?

No, there isn't. BOINC uses redundancy, at least on many projects, where work that is returned in error is sent out to another computer to try. Just to see if it was a faulty computer or a faulty task.
ID: 27940 · Report as offensive
Merciadri Luca

Send message
Joined: 1 Aug 09
Posts: 33
Belgium
Message 27961 - Posted: 13 Oct 2009, 11:56:21 UTC - in response to Message 27940.  

But is there actually an automatic backup for WUs?

No, there isn't. BOINC uses redundancy, at least on many projects, where work that is returned in error is sent out to another computer to try. Just to see if it was a faulty computer or a faulty task.

Thanks for this information.
See http://www.student.montefiore.ulg.ac.be/~merciadri/ for more pieces of info.
ID: 27961 · Report as offensive
rakarin

Send message
Joined: 26 Dec 08
Posts: 9
United States
Message 28059 - Posted: 17 Oct 2009, 3:55:47 UTC - in response to Message 27880.  

Hello,
When I sometimes face OS freezes, or that I encounter electrical problems, my computer needs to be rebooted, just like any other one.


On a tangent... You should have a UPS (battery back-up unit) that can run your PC for at least 60-90 seconds, though three minutes is preferable. I work in IT, and PC's regularly hit with brief power drops and outages develop hardware problems more quickly. Something like an APC 300 or 350 is relatively inexpensive.

Do *not* plug a printer into battery backup. That can draw too much power for the battery and make Bad Things (TM) happen. You want only your PC and monitor on backup, though a network hub / firewall / router / cable modem doesn't draw much power. Leave your speakers and such off the battery backup.

If you leave BOINC (or other distributed computing apps) running at 100%, you will need a larger UPS. Keeping the processor pegged at full eats battery power much more quickly.

Mike
ID: 28059 · Report as offensive
Merciadri Luca

Send message
Joined: 1 Aug 09
Posts: 33
Belgium
Message 28076 - Posted: 17 Oct 2009, 20:15:47 UTC - in response to Message 28059.  


On a tangent... You should have a UPS (battery back-up unit) that can run your PC for at least 60-90 seconds, though three minutes is preferable. I work in IT, and PC's regularly hit with brief power drops and outages develop hardware problems more quickly. Something like an APC 300 or 350 is relatively inexpensive.

I actually have a UPS, but I have so many computers connected to it, that its battery becomes low in a really quick way! I know, I should buy another one.

Another problem is that the computer has to be conscious that it is on batteries, and there are sometimes no USB port, or only one, on the most of the inexpensive UPS. As a result, if I encounter a A/C problem, my computers (at least 2 of the three) are not aware that they are running on batteries, and thus run till the end, thus powering off when batteries are down. What can you do if you are not at home?


Do *not* plug a printer into battery backup. That can draw too much power for the battery and make Bad Things (TM) happen. You want only your PC and monitor on backup, though a network hub / firewall / router / cable modem doesn't draw much power. Leave your speakers and such off the battery backup.

Thanks for your tips. I did not know about the printer.


If you leave BOINC (or other distributed computing apps) running at 100%

I always do this!
, you will need a larger UPS. Keeping the processor pegged at full eats battery power much more quickly.

Thanks.
See http://www.student.montefiore.ulg.ac.be/~merciadri/ for more pieces of info.
ID: 28076 · Report as offensive

Message boards : BOINC client : Computing error

Copyright © 2025 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.