Wierd validations - tasks stuck in limbo?


log in

Advanced search

Message boards : Number crunching : Wierd validations - tasks stuck in limbo?

Author Message
juan BFB
Volunteer tester
Avatar
Send message
Joined: 16 Mar 07
Posts: 4927
Credit: 267,983,673
RAC: 349,525
Brazil
Message 1467638 - Posted: 23 Jan 2014, 12:22:52 UTC
Last modified: 23 Jan 2014, 12:24:06 UTC

I was looking at my old WU (wating for validation or what else for more than 6 months) and see some let´s call "wierd validadted WU", and example:

http://setiathome.berkeley.edu/workunit.php?wuid=1265541696

The first task, was running and exit normaly, the secon one runs for seconds and exit in an overflow, and the third one ends on too many exit error.

The question is why the validator validate work like this?

Is just one of the several old WU some validated and some no for more than a year waiting for? Can´t imagine what some have more than 2 years and still waiting... just a waste space on the DB.

If we are actualy looking for one "golden WU" (the ET call) in the entire millions of WU, then things like this could make us loose the "golden WU".
____________

Profile Wiggo
Avatar
Send message
Joined: 24 Jan 00
Posts: 6465
Credit: 90,203,241
RAC: 73,926
Australia
Message 1467644 - Posted: 23 Jan 2014, 12:35:18 UTC

The date looks to be around the same time that Eric ran his credit granting script which let many of these bad results through.

Cheers.

juan BFB
Volunteer tester
Avatar
Send message
Joined: 16 Mar 07
Posts: 4927
Credit: 267,983,673
RAC: 349,525
Brazil
Message 1467655 - Posted: 23 Jan 2014, 12:56:11 UTC
Last modified: 23 Jan 2014, 13:00:23 UTC

Could be but, then why the DB was not cleared after all this time? Need some time to dig more examples but i´m sure it´s something very common (at least on my hosts).

Or other example compleately diferent example:

http://setiathome.berkeley.edu/workunit.php?wuid=967420872

In this case, the WU were validated and one task still waiting for validation?
Date May/2012. Each each one has few of them will be thousands of WU waiting for what? an "infinite time validation".

or like this: http://setiathome.berkeley.edu/workunit.php?wuid=1252882786 actualy validade but never cleared from the DB for almost a year.

If the DB size is a issue this only could contribute to make the problem even bigger.
____________

Profile Wiggo
Avatar
Send message
Joined: 24 Jan 00
Posts: 6465
Credit: 90,203,241
RAC: 73,926
Australia
Message 1467665 - Posted: 23 Jan 2014, 13:33:19 UTC

Well I have 1 here that goes all the way back to 31 Jul 2010, http://setiathome.berkeley.edu/workunit.php?wuid=967420872 on a rig that I'd like to delete from my list.

I also have on my other 2 rigs;
http://setiathome.berkeley.edu/workunit.php?wuid=1383938907 22 Dec 2013
http://setiathome.berkeley.edu/workunit.php?wuid=1256043119 31 May 2013
http://setiathome.berkeley.edu/workunit.php?wuid=1260822883 8 Jun 2013
http://setiathome.berkeley.edu/workunit.php?wuid=1263335417 13 Jun 2013
http://setiathome.berkeley.edu/workunit.php?wuid=1264963387 16 Jun 2013
http://setiathome.berkeley.edu/workunit.php?wuid=1265755489 18 Jun 2013
http://setiathome.berkeley.edu/workunit.php?wuid=953687194 14 Apr 2012
http://setiathome.berkeley.edu/workunit.php?wuid=1020310310 29 Jun 2012
http://setiathome.berkeley.edu/workunit.php?wuid=1263268374 8 Jul 2013
http://setiathome.berkeley.edu/workunit.php?wuid=1280378195 9 Aug 2013
http://setiathome.berkeley.edu/workunit.php?wuid=1309312251 23 Sep 2013
http://setiathome.berkeley.edu/workunit.php?wuid=1310827896 26 Sep 2013
http://setiathome.berkeley.edu/workunit.php?wuid=1260945737 8 Jun 2013
http://setiathome.berkeley.edu/workunit.php?wuid=1261017996 10 Jun 2013
http://setiathome.berkeley.edu/workunit.php?wuid=1262082648 11 Jun 2013

And then there are another 18 SETI@home Enhanced not included here, but then there are many threads around here on the same subject over a long time to which we have never received an official answer to yet.

Cheers.

Profile HAL9000
Volunteer tester
Avatar
Send message
Joined: 11 Sep 99
Posts: 3849
Credit: 106,558,501
RAC: 92,571
United States
Message 1467666 - Posted: 23 Jan 2014, 13:37:02 UTC

Those in among the tasks stuck in limbo that for whatever reason didn't get cleared out when they should have been. It happened with the previous version of AP & it is now also clear that it happens to MB tasks as well. I have 22 SETI@home Enhanced tasks stuck in suck odd states.

It does seem to be a general issue with the BOINC server which I would not expect to be fixed any time soon.
____________
SETI@home classic workunits: 93,865 CPU time: 863,447 hours

Join the BP6/VP6 User Group today!

juan BFB
Volunteer tester
Avatar
Send message
Joined: 16 Mar 07
Posts: 4927
Credit: 267,983,673
RAC: 349,525
Brazil
Message 1467671 - Posted: 23 Jan 2014, 13:44:42 UTC
Last modified: 23 Jan 2014, 14:01:59 UTC

If DB size is a serius problem as noticed (and the main reason why we still have the 100 GPU WU limit in place) and each one of us has, let said a random number 5, with 100K users we talk about 500K WU doing nothing on DB...

If nothing was made to fix that then something sounds wierd no?

BTW Wiggo i find the issue exactly because i have some host i want to delete from my list as you.

@Hal9000 I like the name you give - added it to the thread title.
____________

Profile HAL9000
Volunteer tester
Avatar
Send message
Joined: 11 Sep 99
Posts: 3849
Credit: 106,558,501
RAC: 92,571
United States
Message 1467684 - Posted: 23 Jan 2014, 14:03:10 UTC

I do not think the issue with the stuck stats is nearly that bad. IIRC when AP v5 ended there were only a few thousand stuck. Which did get cleaned up in time.

However it require Matt or someone in the lab to write the clean up script & run it a few times. It is a bit annoying to us, but from their point of view probably does not effect the project enough to warrant any action. Given that I have stuck tasks that go back as far as 2010 I wouldn't expect it to happen soon. Unless there was a reliable funding source to fund a person to do random clean up like this for them.
____________
SETI@home classic workunits: 93,865 CPU time: 863,447 hours

Join the BP6/VP6 User Group today!

msattler
Volunteer tester
Avatar
Send message
Joined: 9 Jul 00
Posts: 38180
Credit: 556,893,544
RAC: 602,552
United States
Message 1467725 - Posted: 23 Jan 2014, 15:18:22 UTC

I still show 119 Enhanced tasks hanging around in my results.
118 validated and 1 error.
The error task is shown as abandoned. In November of 2010.
____________
*********************************************
Embrace your inner kitty...ya know ya wanna!

I have met a few friends in my life.
Most were cats.

Message boards : Number crunching : Wierd validations - tasks stuck in limbo?

Copyright © 2014 University of California