I upgraded to Boinc 4.13 yesterday morning and everything was running fine. However, about 30 minutes ago, I got the following error message in the log:

climateprediction.net - 2004-10-15 15:22:35 - Unrecoverable error for result 2rsx_200150651_0 ( - exit code -1073741819 (0xc0000005))

Since then it has retrieved 4 units from CPDN and crashed with the same exit code on all 4.

I have CPDN and SETI running with a 20:80 split and SETI@Home is still running happily. I've checked and there is only one client running.

Any ideas?

Tony.

Hi Tony,

It looks like your one of the PC (Host ID 7943) has trouble.

It has shown that almost all the crashes were associated either with underpowered machines and those running Win98/ME/NT (neither of which apply here), or there were factors particular to individual machines which might have been overclocked, have some trouble with memory, problem with graphic card, etc.
Have you run any stability checks on your compute(Host ID 7943)? CPDN does seem to require very stable computers. You may want to read this and try stability checks.
http://www.climateprediction.net/board/viewtopic.php?t=2126

This is only a guess, so it might be a trouble in BOINC 4.13, but I think that chance is low.

> Have you run any stability checks on your compute(Host ID 7943)? CPDN does
> seem to require very stable computers. You may want to read this and try
> stability checks.
> http://www.climateprediction.net/board/viewtopic.php?t=2126

I haven't run stability checks but will do on Monday when I am back at work. It is very strange as this is the most powerful of the computers that I am running it on. I was in a meeting when it failed and the machine was not being used for anything else.

> This is only a guess, so it might be a trouble in BOINC 4.13, but I think that
> chance is low.

I think that this may be possible as it never occured before 4.13. I'll try going back to 4.09 on this machine to see if it makes a difference.

Thanks for the prompt reply,

Tony.

OK. I downloaded the Super PI program and it has been running all morning calculating Pi to various huge numbers of decimal places with no problems at all. Also, when the machine contacted CPDN on Saturday morning, it downloaded a new work unit which it has been running since with no problems.

I also checked in the OS event logs and, apart from the messages on Friday about the failures, there are no other messages. As a last resort, I have searched the computer for any file that was modifed at about the time of the failure and didn't find anything.

Unless it happens again, I'm going to put it down to another mystery of the evil empire and let it carry on with 4.13.

Tony.

Hi Tony,

You did stability checks on your computer and it seems OK, now I don't have any idea why these problems happened.
(How was the temperature of your PC when these problem occured?)

As far as I know, there is no trouble similar to what you experienced on BOINC 4.13, so what I can say is, let's watch your new WU.

Hi Tony,

Maybe this is the same problem, isn't it?
http://climateapps2.oucs.ox.ac.uk/cpdnboinc/forum_thread.php?id=1103

BOINC 4.13 may have something to do with it.

> Maybe this is the same problem, isn't it?
> http://climateapps2.oucs.ox.ac.uk/cpdnboinc/forum_thread.php?id=1103
>
> BOINC 4.13 may have something to do with it.

Hi Arata,

It would certainly appear so. AFter today, I'm going to be out of the office for a week so I'll see how BOINC does on it's own during this time.

I'm still using BOINC 4.13 but the one change that I have made is to set it to leave the paused processes in memory. I'll suggest this in the other thread.

BTW, forgot to say -- temperature was not excessive.

Tony.