What should I do with BOINC 6.2.14?

Vladimir Zarkov
Vladimir Zarkov
Joined: 27 Feb 05
Posts: 66
Credit: 4876895
RAC: 0
Topic 193827

I saw today that it was posted again as a (probably unstable) development version. IT IS unstable, no doubt about it - trashed 3 SETI tasks on my old AMD PC (never happened before), and ALL S5R3 Einstein tasks on my son's Q6600 (there were some compute errors before but never such a toral bust). So do we return to 5.10.45 now, or what?

Mike Hewson
Mike Hewson
Moderator
Joined: 1 Dec 05
Posts: 6588
Credit: 314283196
RAC: 264881

What should I do with BOINC 6.2.14?

Quote:
I saw today that it was posted again as a (probably unstable) development version. IT IS unstable, no doubt about it - trashed 3 SETI tasks on my old AMD PC (never happened before), and ALL S5R3 Einstein tasks on my son's Q6600 (there were some compute errors before but never such a toral bust). So do we return to 5.10.45 now, or what?


I suggest you roll back. 6.2.14 worked on 4 of 5 machines I tried it on - failed on a Dell Inspiron with XP Pro.

Cheers, Mike.

I have made this letter longer than usual because I lack the time to make it shorter ...

... and my other CPU is a Ryzen 5950X :-) Blaise Pascal

Jord
Joined: 26 Jan 05
Posts: 2952
Credit: 5893653
RAC: 137

First things first, since

First things first, since BOINC doesn't crunch, how can it be trashing tasks? It manages the project applications. Yes, by not installing it correctly can you make it go haywire, but just following the default setup, should be a walk-in-the-park.

Quote:
trashed 3 SETI tasks on my old AMD PC (never happened before)


They don't show in your list of tasks at Seti. So what exactly was their error message then?

Quote:
and ALL S5R3 Einstein tasks on my son's Q6600 (there were some compute errors before but never such a toral bust).


That machine first had an exit code 10, where for some reason the file name of the task changed names all of a sudden:

Checksum error: 3519315
Could not resume from checkpoint (-2)
ERROR: MAIN() returned with error '10'
zip warning: name not matched: h1_1196.75_S5R3__54_S5R3b_1_0
WARNING: Can't zip output file 'h1_1196.75_S5R3__54_S5R3b_1_0'


Errors like this one signal there's something wrong with the drive. Either the controller is going, or the sector that was written to is damaged. Run chkdsk /F on that drive/partition.

Then with 6.2.14 you started to return "CreateProcess() failed - Access is denied. (0x5)" errors only. See this thread for a very possible cause: Another program, like the anti virus program or a firewall, is stopping BOINC from starting the application.

Vladimir Zarkov
Vladimir Zarkov
Joined: 27 Feb 05
Posts: 66
Credit: 4876895
RAC: 0

Thanks, Mike and Jord. Just

Thanks, Mike and Jord. Just to be on the safe side I'll use 5.10.45 again.

I don't know why those trashed SETI tasks don't show anymore in my account there. The server must have cleaned them up, as the compute errors happened some days ago.
This is the only info I found in stdoutdae.txt (I'm not sure where to look):

[08/03/08 12:45:24] TRACE [2940]: Event: CTRL-LOGOFF Event

[08/03/08 12:45:38] TRACE [3136]: ***** Console Event Detected *****

[08/03/08 12:45:38] TRACE [3136]: Event: CTRL-LOGOFF Event

[08/03/08 12:45:42] TRACE [3264]: ***** Console Event Detected *****

[08/03/08 12:45:42] TRACE [3264]: Event: CTRL-CLOSE or CTRL-SHUTDOWN Event

I can't guess what happened - may be the Comodo firewall didn't like what the new BOINC was doing, and stopped it by brute force but this would be the first problem with BOINC in all the time I used Comodo.
May be the same happened with the Q6600 machine. (The same set there - NOD32 and Comodo.) Or may be not. It's a gamer's rig where anything could happen but before 6.2.14 things were going smoothly, more or less.

Thanks again for your attention and time, which I would not like to waste. I'm a barely adequate user, so the best and simplest solution for me is to go back to the version without problems.
Have a nice day!

Jord
Joined: 26 Jan 05
Posts: 2952
Credit: 5893653
RAC: 137

RE: may be the Comodo

Message 84084 in response to message 84083

Quote:
may be the Comodo firewall didn't like what the new BOINC was doing, and stopped it by brute force but this would be the first problem with BOINC in all the time I used Comodo.


Then check the other thread I linked to. That person is also using Comodo and he's back at crunching with 6.2.14 after he told the firewall that BOINC and its applications were OK (again). That's more of a bug in Comodo, that it doesn't pop-up a new request to allow a new or updated program through, than it is a problem with BOINC.

Besides, you have the possibility that it will do this now with 5.10.45 as well as the firewall has detected that the program changed. What do you think it'll do when you change the program again? ;-)

So just open Comodo and re-allow BOINC to execute programs and access the internet. That's all there is to it.

Vladimir Zarkov
Vladimir Zarkov
Joined: 27 Feb 05
Posts: 66
Credit: 4876895
RAC: 0

Damn... The problem was both

Damn... The problem was both with SETI and BOINC 6.2.14. The new version moved the application folders to Documents and Settings, and Comodo interpreted it as a need for new settings in Defense+/Advanced/Computer Security Policy. Custom settings at that. Everything was set to Ask, that is, the firewall stopped the process, and the user should expressly allow it to go on. But no pop-up message. No wonder all those tasks went to waste with compute errors. It only added to the mess - I've allowed other security programs to make some tweaks which I had to cancel and start again... It looks all right now though.
Thanks again for your help!

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.