climateprediction.net home page
IS

IS

Message boards : Number crunching : IS
Message board moderation

To post messages, you must log in.

AuthorMessage
Arnaud

Send message
Joined: 3 Sep 04
Posts: 268
Credit: 256,045
RAC: 0
Message 8738 - Posted: 6 Feb 2005, 13:13:49 UTC

Hi,

1.Is it possible to continue to crunch a Wu (thanks to a back-up) when the server status says the WU has crashed (unrecoverable error code -5)
2.Will the trickle continue althought the error status ?
3.Is it possible to upload the results (manually or with another method)
Or if the WU crashed, the game is over and I have to start another Wu ?

My Wu didn't crash but I have these questions frequently on the french forum, and usually I answer that the Wu is lost when the server status says "unrecoverable error" or "- exit code -5 (0xfffffffb) ". IS it really true ? IS the wu really lost or can we continue to crunch it without paying attention to the server messages ?
ID: 8738 · Report as offensive     Reply Quote
crandles
Volunteer moderator

Send message
Joined: 16 Oct 04
Posts: 692
Credit: 277,679
RAC: 0
Message 8741 - Posted: 6 Feb 2005, 13:38:47 UTC - in response to Message 8738.  

> Hi,
>
> 1.Is it possible to continue to crunch a Wu (thanks to a back-up) when the
> server status says the WU has crashed (unrecoverable error code -5)
> 2.Will the trickle continue althought the error status ?
> 3.Is it possible to upload the results (manually or with another method)
> Or if the WU crashed, the game is over and I have to start another Wu ?
>
> My Wu didn't crash but I have these questions frequently on the french forum,
> and usually I answer that the Wu is lost when the server status says
> "unrecoverable error" or "- exit code -5 (0xfffffffb) ". IS it really true ?
> IS the wu really lost or can we continue to crunch it without paying attention
> to the server messages ?
>
>
Several people have said that they have had no problem. Just backup whole boinc directory and after crash restore the directory. It gets a lot more complicated if there are other projects and you do not want to loose the work on those other projects since the last backup.

You get credit for trickles once you pass the last trickle submitted.

I do not think the server status matters much. There is a science database and a database for trickles etc. The science database is what matters to the scientists so they are not going to let a problem with the trickle database prevent upload of valid data.
Visit BOINC WIKI for help

And join BOINC Synergy for all the news in one place.
ID: 8741 · Report as offensive     Reply Quote
Holmis

Send message
Joined: 31 Aug 04
Posts: 17
Credit: 303,467
RAC: 0
Message 8742 - Posted: 6 Feb 2005, 13:40:21 UTC - in response to Message 8738.  

> Hi,
>
> 1.Is it possible to continue to crunch a Wu (thanks to a back-up) when the
> server status says the WU has crashed (unrecoverable error code -5)
> 2.Will the trickle continue althought the error status ?
> 3.Is it possible to upload the results (manually or with another method)
> Or if the WU crashed, the game is over and I have to start another Wu ?
>
> My Wu didn't crash but I have these questions frequently on the french forum,
> and usually I answer that the Wu is lost when the server status says
> "unrecoverable error" or "- exit code -5 (0xfffffffb) ". IS it really true ?
> IS the wu really lost or can we continue to crunch it without paying attention
> to the server messages ?

Can't remember who said it but it's not over, restore a full copy of the boinc-folder and the server will catch up when a new trickle is recieved. If the trickle is already send before nothing should happen and the first that is new should get credit.

What happens when the result is completed and uploaded I can't say.
ID: 8742 · Report as offensive     Reply Quote
Arnaud

Send message
Joined: 3 Sep 04
Posts: 268
Credit: 256,045
RAC: 0
Message 8744 - Posted: 6 Feb 2005, 14:32:53 UTC

Thanks a lot for your answers :o)
ID: 8744 · Report as offensive     Reply Quote
old_user23880
Volunteer tester

Send message
Joined: 10 Oct 04
Posts: 223
Credit: 4,664
RAC: 0
Message 8777 - Posted: 7 Feb 2005, 4:09:43 UTC

But didn't Carl say a few months back that error code -5 often meant that negative pressure values had been generated? If this is the case,and there has been a calculation error, should users really try to save these models?

I had several models abort with code -5, and they were invariably already developing crazy weather by 1812/13, when the climate should be stable. So I made no attempt to save and recover.
__________________________________________________

ID: 8777 · Report as offensive     Reply Quote
Arnaud

Send message
Joined: 3 Sep 04
Posts: 268
Credit: 256,045
RAC: 0
Message 8780 - Posted: 7 Feb 2005, 6:55:19 UTC
Last modified: 7 Feb 2005, 7:01:51 UTC

Well, presently I see cases of -5 code after upgrading CC4.13 toward CC4.19 and I don't think it's a negative pression problem.

I had more than 50 unrecoverable errors -5 since I began CPDN and I'm sure it wasn't 50 instable models.
In fact this kind of errors come from instable machines, dumb users like me who know nothing about computer except click on the mouse, and of BOINC not being always very stable.
So I make all kind of attempt to save and recover :o)

ID: 8780 · Report as offensive     Reply Quote
old_user1132

Send message
Joined: 25 Aug 04
Posts: 28
Credit: 6,522,252
RAC: 0
Message 8795 - Posted: 7 Feb 2005, 10:56:58 UTC - in response to Message 8777.  

> But didn't Carl say a few months back that error code -5 often meant that
> negative pressure values had been generated? If this is the case,and there has
> been a calculation error, should users really try to save these models?
>
> I had several models abort with code -5, and they were invariably already
> developing crazy weather by 1812/13, when the climate should be stable. So I
> made no attempt to save and recover.
>

Mo,
I always try and retrieve the model at least once, if it's a genuine model/parameter space error it will fail again and I give up. Most of my problems have been Windoze throwing a wobbly or more often still my finger trouble..

Andrew
Andrew

<a href="http://cpdnforum.info">CPDNforum<a>
ID: 8795 · Report as offensive     Reply Quote
old_user1
Avatar

Send message
Joined: 5 Aug 04
Posts: 907
Credit: 299,864
RAC: 0
Message 8873 - Posted: 8 Feb 2005, 1:40:15 UTC - in response to Message 8795.  

-5 is a "catch all" -- the negative pressure is pretty rare (although it was popping up a lot when I was building the model cross-platform early in the beta test).
ID: 8873 · Report as offensive     Reply Quote

Message boards : Number crunching : IS

©2024 cpdn.org