Advanced search

Forums : Technical Support : Result not needed?
Message board moderation

To post messages, you must log in.

AuthorMessage
sygopet

Send message
Joined: 2 Aug 08
Posts: 27
Credit: 204,771
RAC: 0
Message 6999 - Posted: 11 Aug 2008, 11:09:26 UTC

As a new cruncher with this project I find a lot of disatisfaction on the message boards but not anything which quite corresponds to my experience indicated in the following extract:

11/08/2008 10:27:10|Cosmology@Home|Sending scheduler request: To fetch work. Requesting 11880 seconds of work, reporting 2 completed tasks
11/08/2008 10:27:17|Cosmology@Home|Scheduler request succeeded: got 0 new tasks
11/08/2008 10:27:17|Cosmology@Home|Message from server: Completed result wu_072508_020344_1_1 refused: this result wasn\'t sent (not needed)
11/08/2008 10:27:17|Cosmology@Home|Message from server: Completed result wu_072508_064112_0_3 refused: this result wasn\'t sent (not needed)
11/08/2008 10:27:17|Cosmology@Home|Message from server: No work sent

If anyone can check this and explain, I would be grateful.
Also, if this duplicates a problem discussed elsewhere (which I haven\'t found) moderators may feel free to edit or transfer this posting to the right thread!
ID: 6999 · Report as offensive     Reply Quote
STE\/E
Volunteer tester

Send message
Joined: 12 Jun 07
Posts: 375
Credit: 16,539,257
RAC: 0
Message 7000 - Posted: 11 Aug 2008, 11:32:05 UTC
Last modified: 11 Aug 2008, 11:34:08 UTC

Your @ the Cosmology Project, that should explain it all really ... ;)

Actually it probably has something to do with this Post made by the Project on the 9\'th of Aug ...

August 9, 2008 Old WUs cancalled and feeder code modified
Workunits created over two weeks ago have been piling up, so we\'ve made a small change to the feeder code and cancelled the offending WUs.


Your Wu\'s got Canceled because they were Offending the Project if I read that Post right ... LOL
ID: 7000 · Report as offensive     Reply Quote
Scrooby

Send message
Joined: 13 Feb 08
Posts: 77
Credit: 616,140
RAC: 0
Message 7001 - Posted: 11 Aug 2008, 18:45:14 UTC - in response to Message 6999.  

As a new cruncher with this project I find a lot of disatisfaction on the message boards but not anything which quite corresponds to my experience indicated in the following extract:

11/08/2008 10:27:17|Cosmology@Home|Message from server: Completed result wu_072508_064112_0_3 refused: this result wasn\'t sent (not needed)
11/08/2008 10:27:17|Cosmology@Home|Message from server: No work sent



Its really still a Beta project, this sort of thing happens.
ID: 7001 · Report as offensive     Reply Quote
Brian Silvers

Send message
Joined: 11 Dec 07
Posts: 420
Credit: 270,580
RAC: 0
Message 7002 - Posted: 11 Aug 2008, 19:03:43 UTC - in response to Message 7001.  

As a new cruncher with this project I find a lot of disatisfaction on the message boards but not anything which quite corresponds to my experience indicated in the following extract:

11/08/2008 10:27:17|Cosmology@Home|Message from server: Completed result wu_072508_064112_0_3 refused: this result wasn\'t sent (not needed)
11/08/2008 10:27:17|Cosmology@Home|Message from server: No work sent



Its really still a Beta project, this sort of thing happens.


\"This is just a beta project\" should not be used as justification for releasing a mostly-untested application out to the entire user base. That is not the traditional definition of \"beta\", but more along the lines of \"alpha\".


ID: 7002 · Report as offensive     Reply Quote
Scrooby

Send message
Joined: 13 Feb 08
Posts: 77
Credit: 616,140
RAC: 0
Message 7003 - Posted: 11 Aug 2008, 20:14:25 UTC - in response to Message 7002.  
Last modified: 11 Aug 2008, 20:25:22 UTC

As a new cruncher with this project I find a lot of disatisfaction on the message boards but not anything which quite corresponds to my experience indicated in the following extract:

11/08/2008 10:27:17|Cosmology@Home|Message from server: Completed result wu_072508_064112_0_3 refused: this result wasn\'t sent (not needed)
11/08/2008 10:27:17|Cosmology@Home|Message from server: No work sent



Its really still a Beta project, this sort of thing happens.


\"This is just a beta project\" should not be used as justification for releasing a mostly-untested application out to the entire user base. That is not the traditional definition of \"beta\", but more along the lines of \"alpha\".



If you sign up here, you\'re a tester. If you dont want to see work get lost or trashed, go away for a while.


# Grant the full 70 credits to any task that was returned with \"success\" for the past 3 weeks.

# If this project is now going to take a stance of \"it\'s \'beta\', so if there are problems, expect 100% loss of your time and energy\", then it should state that motto somewhere on the front page of the project.


Problem with all these projects is nobody has enough time to do stuff like that.
As we\'ve discovered, Ben has lost his fulltime assistant at a critical time.

FWIW I did a whole bunch of 100+hr tasks at SETI Beta, they all got flagged \"not needed\".
ID: 7003 · Report as offensive     Reply Quote
Profile Jord
Volunteer moderator
Volunteer tester
Avatar

Send message
Joined: 15 Jun 07
Posts: 345
Credit: 50,500
RAC: 0
Message 7004 - Posted: 11 Aug 2008, 20:29:06 UTC - in response to Message 7002.  

\"This is just a beta project\" should not be used as justification for releasing a mostly-untested application out to the entire user base. That is not the traditional definition of \"beta\", but more along the lines of \"alpha\".

Alpha is setting up the project, bring everything on-line, see if your first experimental applications even like to be on your computers, hope you can get your back-end up and see things crash and burn.

Normally Alpha is run internal. You can ask a small group of people to come test the software for you, although this isn\'t necessary.

Beta is testing your latest applications based on stable code, you hope your back-end stays up&running, you expect to get new bugs, you\'ve asked a larger group of people on the outside world to come test your code and applications. You want feedback. You still expect some crashes and burning. based upon the reports you get you release different applications, see which ones work and which ones definitely don\'t.
ID: 7004 · Report as offensive     Reply Quote
Profile Ananas

Send message
Joined: 19 Jan 08
Posts: 180
Credit: 2,500,290
RAC: 0
Message 7005 - Posted: 11 Aug 2008, 21:56:19 UTC
Last modified: 11 Aug 2008, 22:38:44 UTC

Just a few thoughts ...

If there had been a project preference \"Enable alpha workunits\", disabled by default - and while testing, only alpha workunits would have been available - how many would have enabled it when the credits have been really high?

I bet, nearly everyone of us would have done that.


The project application is not unstable at all, it doesn\'t crash and doesn\'t make the hosts crash, so it is much better than alpha.


The web server problems are nothing that can be tested so easily in a standalone setup, that can only be tested under production conditions.

Something must be weird in the setup required by BOINC, a lot of projects fall into the \"no work sent\" trap at least once in their startup time. There should be a self-test script for BOINC that checks all requirements and tells the project admin about potential problems and traps.

Without those server problems, the step to cancel / invalidate old results and retry them with a new program version would not have hit such a high amount of results.

Beeing mad when something fails is understandable, but I think it\'s getting a bit out of hands here at the moment. A very small team handling a complete BOINC project with limited ressources, we just cannot expect it to run as perfect as a project with 5 developers, 3 admins and a huge monthly budged.


p.s.: That \"not needed\" BOINC bug is something that currently has hit LHC too, crunchers are not happy there either - which means that it isn\'t Cosmology\'s fault that it is handled in such a crappy way.
ID: 7005 · Report as offensive     Reply Quote
Brian Silvers

Send message
Joined: 11 Dec 07
Posts: 420
Credit: 270,580
RAC: 0
Message 7006 - Posted: 11 Aug 2008, 22:46:14 UTC - in response to Message 7005.  


The project application is not unstable at all, it doesn\'t crash and doesn\'t make the hosts crash, so it is much better than alpha.


This is almost like dealing with Ned over at SETI...

[maximum verbosity]
The initial Linux apps that had the lensing re-enabled were definitely of \"alpha\" quality, as they crashed immediately on a large number of hosts. During the rush to get this fixed, the back-end was clearly not tested either, as cross-application validation did not work properly, as Windows application version 2.12-processed tasks would not cross-validate with Windows application version 2.14. Windows application version 2.14 tasks did cross-validate with Windows application version 2.15, but the project attempted to rid themselves of anything other than version 2.15, while the real problem was just with the hosts dropping out because of problems with versions 2.12 and 2.13 that could\'ve been addressed by doing internal testing before releasing those application versions or by creating a test application utilizing the Anonymous Platform mechanism.
[/maximum verbosity]

I AM a software developer. Given the amount of problems with the whole overall process, what we went through was Alpha. In theory, now that things are supposedly more stable with the applications and the back-end, now we are at Beta.

I should say \"supposedly\", because I think the effort to get rid of the \"old\" results was only partially successful. I think we\'ll still be seeing them get regenerated as they time out.


Beeing mad when something fails is understandable, but I think it\'s getting a bit out of hands here at the moment. A very small team handling a complete BOINC project with limited ressources, we just cannot expect it to run as perfect as a project with 5 developers, 3 admins and a huge monthly budged.


I am not asking for \"perfection\". I\'m asking for \"due diligence\" on their part. They knew the staffing situation prior to re-enabling lensing. If they were not able to test the application due to low staffing levels, then they should\'ve been prepared for a higher degree of support. If they were not prepared for a higher level of support due to the low staffing levels, which clearly they were and are not, then the process should\'ve waited until such time as they had either more testing or support resources.



ID: 7006 · Report as offensive     Reply Quote
Brian Silvers

Send message
Joined: 11 Dec 07
Posts: 420
Credit: 270,580
RAC: 0
Message 7007 - Posted: 11 Aug 2008, 22:51:04 UTC - in response to Message 7005.  


p.s.: That \"not needed\" BOINC bug is something that currently has hit LHC too, crunchers are not happy there either - which means that it isn\'t Cosmology\'s fault that it is handled in such a crappy way.


I don\'t know how one can tell what condition a workunit is in at LHC to be able to know if one should\'ve gotten a cancellation based on the quorum being met, considering all attempts to look at a workunit over there lead to the following:


Fatal error: Call to undefined function get_int() in /srv/httpd/lhcathome.cern.ch/lhcathome/html/user/workunit.php on line 8



ID: 7007 · Report as offensive     Reply Quote
Scrooby

Send message
Joined: 13 Feb 08
Posts: 77
Credit: 616,140
RAC: 0
Message 7008 - Posted: 11 Aug 2008, 23:31:18 UTC - in response to Message 7006.  

I am not asking for \"perfection\". I\'m asking for \"due diligence\" on their part.


At least you\'ve stopped asking for your RAC to be restored.

ID: 7008 · Report as offensive     Reply Quote
Brian Silvers

Send message
Joined: 11 Dec 07
Posts: 420
Credit: 270,580
RAC: 0
Message 7010 - Posted: 12 Aug 2008, 2:22:34 UTC - in response to Message 7008.  
Last modified: 12 Aug 2008, 3:19:50 UTC

I am not asking for \"perfection\". I\'m asking for \"due diligence\" on their part.


At least you\'ve stopped asking for your RAC to be restored.



Psst... I\'m not all that greatly affected... I was asking for the people who were/are greatly affected...

Since we\'re on the subject, the project did something that caused the loss of processing time uneccessarily. They sledgehammered the problem. The sad thing is, they didn\'t \"fix it\". Old tasks are still being worked on. Instead of investigating the situation, they appear to have went all:

Leeroy Jenkins
Youtube video of Leeroy

Just how many mega-watts of power were consumed needlessly for that? How many more will be consumed to recover from it?
ID: 7010 · Report as offensive     Reply Quote
Profile speedimic
Avatar

Send message
Joined: 9 Sep 07
Posts: 89
Credit: 2,201,260
RAC: 0
Message 7011 - Posted: 12 Aug 2008, 16:06:08 UTC

[quote]...Youtube video of Leeroy[/qoute]

Cool, haven\'t seen that one.

Leeeeeerrooooooyyyy!!!
mic.


ID: 7011 · Report as offensive     Reply Quote

Forums : Technical Support : Result not needed?