Weird redundancy
log in

Advanced search

Message boards : Number crunching : Weird redundancy

Author Message
Profile valterc
Project administrator
Project tester
Send message
Joined: 30 Oct 13
Posts: 616
Credit: 34,514,943
RAC: 395
Italy
Message 64 - Posted: 18 Dec 2013, 10:13:16 UTC

I noticed that some of the workunits crunched by boboviz are in "validation inconclusive" state. See here http://gene.disi.unitn.it/test/results.php?hostid=16&offset=0&show_names=0&state=3&appid=.

Workunit # 21011 has already three completed results that don't agree each other. This could be a problem, please investigate...

Also Workunit # 21087 has two results but the workunit has been sent to the same computer... So boboviz is the wingman of himself (and he don't agree with himself, which is something that could also happens in the real life...). But this shouldn't happen here.....

Profile [VENETO] boboviz
Send message
Joined: 12 Dec 13
Posts: 182
Credit: 4,633,870
RAC: 28
Italy
Message 66 - Posted: 18 Dec 2013, 11:00:52 UTC - in response to Message 64.

So boboviz is the wingman of himself (and he don't agree with himself, which is something that could also happens in the real life...)


:-)

Profile valterc
Project administrator
Project tester
Send message
Joined: 30 Oct 13
Posts: 616
Credit: 34,514,943
RAC: 395
Italy
Message 77 - Posted: 18 Dec 2013, 17:59:21 UTC - in response to Message 66.

I guess I fixed it modifying the following lines inside config.xml

1
10
50


Is there anyone from the 'server' group listening?

chau
Send message
Joined: 12 Nov 13
Posts: 15
Credit: 229
RAC: 0
Italy
Message 85 - Posted: 19 Dec 2013, 12:49:18 UTC - in response to Message 77.
Last modified: 19 Dec 2013, 13:00:16 UTC

Hi Valter,

There are somethings wrong with the results from client. I also have the same situation, I have 2 tasks of WU 21058 running on my computer. The results are already completed and sent back to servers but when I check it, they have different results and different check_sum. That's why the validator marked as inconclusive.

Profile paolomorettin
Project developer
Project tester
Project scientist
Send message
Joined: 20 Nov 13
Posts: 19
Credit: 13,027
RAC: 0
Message 86 - Posted: 19 Dec 2013, 13:36:10 UTC

we're checking it right now :)
____________
Paolo - Application team dev (SSC11)

"If you were plowing a field, which would you rather use: two strong oxen or 1024 chickens?" Seymour Cray

Profile paolomorettin
Project developer
Project tester
Project scientist
Send message
Joined: 20 Nov 13
Posts: 19
Credit: 13,027
RAC: 0
Message 99 - Posted: 19 Dec 2013, 18:22:36 UTC

I made a thread on the development board:

http://gene.disi.unitn.it/test/forum_thread.php?id=27#98
____________
Paolo - Application team dev (SSC11)

"If you were plowing a field, which would you rather use: two strong oxen or 1024 chickens?" Seymour Cray

Profile valterc
Project administrator
Project tester
Send message
Joined: 30 Oct 13
Posts: 616
Credit: 34,514,943
RAC: 395
Italy
Message 100 - Posted: 19 Dec 2013, 18:35:37 UTC - in response to Message 85.
Last modified: 19 Dec 2013, 18:36:38 UTC

Hi Valter,

There are somethings wrong with the results from client. I also have the same situation, I have 2 tasks of WU 21058 running on my computer.

There should be no more this problem. When a new workunit is generated the two copies are no more given to the same user. This applies only to the workunits generated *after* yesterday in the evening.

Profile [VENETO] boboviz
Send message
Joined: 12 Dec 13
Posts: 182
Credit: 4,633,870
RAC: 28
Italy
Message 148 - Posted: 23 Dec 2013, 8:45:34 UTC - in response to Message 64.

I noticed that some of the workunits crunched by boboviz are in "validation inconclusive" state.


A lot of validation inconclusive... (125 wus)
:-(

Profile valterc
Project administrator
Project tester
Send message
Joined: 30 Oct 13
Posts: 616
Credit: 34,514,943
RAC: 395
Italy
Message 149 - Posted: 23 Dec 2013, 11:40:41 UTC - in response to Message 148.

most of them will be eventually marked as invalid due to a bug in the first version of the application. We may hack a little bit the db to give credits anyway..


Post to thread

Message boards : Number crunching : Weird redundancy


Main page · Your account · Message boards


Copyright © 2024 CNR-TN & UniTN