climateprediction.net home page
Problem with stats - results info

Problem with stats - results info

Questions and Answers : Windows : Problem with stats - results info
Message board moderation

To post messages, you must log in.

AuthorMessage
old_user90015

Send message
Joined: 26 Jul 05
Posts: 2
Credit: 1,431,351
RAC: 0
Message 27374 - Posted: 17 Mar 2007, 17:08:27 UTC

Some time back I had a crash that corrupted the 2 projects running on my dual core desktop computer, but fortunately I had backed it up and restarted from the backup. However under \"results\" it says \"over\" under server state and \"client error\" under outcome and \"done\" under client state. This is in spite of receiving trickles ever since and me getting credit for the trickles.

Now that the 2 projects are getting close to 100% I\'m getting worried that the results will be discarded or not accepted because of the \"client error\", so I wondered if the two projects could be reset to their actual active state (of course without me loosing the credits and progress!!). The 2 projects are hadcm3lbm_amrh_25265328_1 and hadcm3lbm_8le5_25170240_0

Thanks.
ID: 27374 · Report as offensive     Reply Quote
Les Bayliss
Volunteer moderator

Send message
Joined: 5 Sep 04
Posts: 7629
Credit: 24,240,330
RAC: 0
Message 27375 - Posted: 17 Mar 2007, 17:17:13 UTC

Nice to see that you\'re making backups. :)

Because none of the many other BOINC projects need or use backups, the server software has no provision for changing the labels when new ones are received.
So the first ones to get to the server stay.

However, none of the credits are lost, and all of the data sent back is saved and used, as the \'research\' side doesn\'t use the messages/labels.

And when you finally finish, you\'ll get a message rejecting your results because of the previous error.
Again, ignore this. It too is just for the other projects and doesn\'t get used.

Just in case you haven\'t seen them, there\'s help, hints, tips, and advice in the 4 READMEs here.

ID: 27375 · Report as offensive     Reply Quote
old_user90015

Send message
Joined: 26 Jul 05
Posts: 2
Credit: 1,431,351
RAC: 0
Message 27376 - Posted: 17 Mar 2007, 17:22:39 UTC - in response to Message 27375.  


However, none of the credits are lost, and all of the data sent back is saved and used, as the \'research\' side doesn\'t use the messages/labels.


Great, that was my biggest worry. I don\'t really mind about the credits, just as long as the research data isn\'t lost and the CPU use hasn\'t been for nothing. :)


And when you finally finish, you\'ll get a message rejecting your results because of the previous error.
Again, ignore this. It too is just for the other projects and doesn\'t get used.


OK, good to know. I would have been crushed to receive that message without knowing this.

Thanks again.
ID: 27376 · Report as offensive     Reply Quote
Profile old_user17289

Send message
Joined: 13 Sep 04
Posts: 228
Credit: 354,979
RAC: 0
Message 27389 - Posted: 19 Mar 2007, 5:16:24 UTC

I too had in the past crashed models that I was able to recover from backups. The CPDN website shows the model status as \'Client Error\' until the final result is uploaded, then it is changed to \'Success\'. All error information disappear after the final upload.
ID: 27389 · Report as offensive     Reply Quote
Les Bayliss
Volunteer moderator

Send message
Joined: 5 Sep 04
Posts: 7629
Credit: 24,240,330
RAC: 0
Message 27390 - Posted: 19 Mar 2007, 5:23:15 UTC
Last modified: 19 Mar 2007, 5:29:14 UTC

That\'s interesting.
There must have been a change in one of the software upgrades over the last year or so.

On the other hand ...
I\'ve just checked my Results page, and the model which I\'ve been trying to recover for the past week shows Over Success Done, even though I had to finally let it go after a 3rd failure at the same point.
And it wasn\'t quite halfway.

ID: 27390 · Report as offensive     Reply Quote
Profile Strathpeffer
Avatar

Send message
Joined: 9 Jan 07
Posts: 497
Credit: 342,899
RAC: 0
Message 27395 - Posted: 19 Mar 2007, 15:54:11 UTC

I and quite a few other team Scotland members have had whole strings of \"error messages\" with models that have been accepted as completed. Some of those errors show throughout and some appear only when the final result is downloaded.

Also someone in team Wales thinks he generates a \"not a JPEG file\" error every time he selects \"view graphics\"!

None of this seems to matter, so long as your model keeps trickling to the end of 2080! (And in any case, the project is now considering models that get past 2050 to be complete - but they would still prefer us to go on to 2080 and there\'s somehow more satisfaction in doing that!)

Good luck!
Visit the Scotland team
ID: 27395 · Report as offensive     Reply Quote

Questions and Answers : Windows : Problem with stats - results info

©2024 climateprediction.net