Index  | Recent Threads  | Unanswered Threads  | Who's Active  | Guidelines  | Search
 

Quick Go »
No member browsing this thread
Thread Status: Active
Total posts in this thread: 2
[ Jump to Last Post ]
Post new Thread
Author
Previous Thread This topic has been viewed 1519 times and has 1 reply Next Thread
SekeRob
Master Cruncher
Joined: Jan 7, 2013
Post Count: 2741
Status: Offline
Reply to this Post  Reply with Quote 
Are the Techs confident ALL results are being indexed for statistics?

I'm not.

Day after day, kind of since moving the cloud, I see more results validated than show up in the statics, not 1, not 2, sometimes 50 - 60 and it seems to focus on FAHV.

Since it's always shortfall, rarely additional, and then mostly 1, these cases are easily identifiable by the last mod timestamp being right on the brink of 00:06:02/12:06:02 or thereabouts.

Yesterday noon, 55 missing for 0:06:05:07, i.e. I got them, but the stats have not, and by exception showed WCG having a shortfall of 1 SCC at 02:06:17 hours returned 5-8-2017 07:22:30, validated at 17-08-12 23:14:11, not anywhere close to the spill-time cutoff. (SCC1_0000774_Bct-E_14333_1 ResultID 1680560493).

The other 54 were for FAHV, but if there are 53.5K in the run and dozens to the second exact the same, there's no telling, but the API flags being perfectly identical, and having validated been between 12:06:03 and 00:06:02 last nite, UTC of course.

My prog eliminates duplicates based on ResultId/WorkUnitID uniqueness, at least I'm hoping we have true uniqueness for these keys where again logic would in that case dictate, to then be left with less than what WCG includes in the stats, not more.... highly improbable. When archiving the valid lot, including the overage, again the keys are checked, and again the pop up says, "0 duplicates removed".

WCG having more is logical, as the pass could have missed them during an API fetch run, WCG having less is not, particularly, if the previous stats cycle showed a perfect match.

What am I missing (the famous question is). Blame MS, at which point there's a world wide business problem ;?
[Aug 13, 2017 1:37:10 PM]   Link   Report threatening or abusive post: please login first  Go to top 
knreed
Former World Community Grid Tech
Joined: Nov 8, 2004
Post Count: 4504
Status: Offline
Project Badges:
Reply to this Post  Reply with Quote 
Re: Are the Techs confident ALL results are being indexed for statistics?

The BOINC system doesn't keep track of the number of results that a user completes. That is data we created when we exported the stats.

However, you could check to see if there is something wrong with the export process since we export credit/points along with count of results and if one were wrong, the other should be wrong also.

This query:

https://www.worldcommunitygrid.org/boinc/show_user.php?userid=1158

wilil give you the current number of credits a user has earned (in this case it is my admin account). This is constantly updated as validation is happening.

This query:

https://www.worldcommunitygrid.org/stat/viewMemberInfo.do?userName=knreed

will give you the totals stats accumulated through the last time that the stats ran for the website.

With a couple of gotchas:

The website includes both points accumulated with United Devices clients.
The websites points are 7 times the BOINC credits.

You should be able to query the BOINC credit at 0:01 UTC each day and the credit gained there (multiplied by 7) should be similar to the daily points reported by the website ( adjusted for rounding). I would track this over 5-6 days and see if they are running similar.

This will give you a way to investigate further and allow it to work as a check on the other stuff you are doing in your script.
[Aug 14, 2017 2:21:54 PM]   Link   Report threatening or abusive post: please login first  Go to top 
[ Jump to Last Post ]
Post new Thread