| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Community
Forum: Teams
Thread: UK |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 14623
|
|
| Author |
|
|
dcrobinson
Veteran Cruncher UK Joined: Mar 10, 2009 Post Count: 1176 Status: Offline Project Badges: 
|
Good morning!
----------------------------------------My stats run failed last night. I've had a quick look at it, and it looks like an issue with the data on the WCG servers. It's been pulling down thousands of data records every day for 11 months with virtually no problems, but it's now giving an error that indicates that some of the data is missing (for one of the team members). This breaks the code that shoves everything into the database. This is not something I can fix in 10 minutes before work after a bad night's sleep. I'm hoping it will mend itself in the lunchtime stats update, so I'll run it again this afternoon. Otherwise, it could be a few days before I can fix it. Sorry folks.
Dave Robinson, Malvern, UK
|
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Good morning everyone.
 Dave, I think it is "mea culpa", as part the shenanigans with over the last couple of weeks on my PC, including the boot recovery. My machine used to be called Douglas_System_Device. 1st Recovery attempt I chose to rename it Twoflower, and the second attempt, the one that is now running smoothly (so far it seems) I had also named the device Twoflower.  Looking at My Device Statistics I find the following BOINC Twoflower 10/12/2016 04:42:43 0:034:12:36:19 150,264 406 BOINC android 10/12/2016 00:54:46 1:285:09:57:07 479,029 1,718 BOINC Twoflower 10/03/2016 12:47:23 5:329:19:05:30 10,794,249 28,600 BOINC Librarian 09/15/2015 21:38:23 2:363:12:38:15 4,203,882 14,386 As you can see a complete mess. The first instance of Twoflower is the current machine, the second instance Twoflower appears to be the combination of Douglas-System-Device and initial attempt of rebuild. BOINC android is what it is and no change, the device called Librarian is on the Win7 partition and no longer used. |
||
|
|
dcrobinson
Veteran Cruncher UK Joined: Mar 10, 2009 Post Count: 1176 Status: Offline Project Badges:
|
I don't think you're the problem domonijo. I've had lots of machines with the same name over the years, and have moved BOINC folders between them.
----------------------------------------For each team member, there's a block of XML data returned from the website, which is supposed to follow a certain structure, and my software decodes that against a "schema" (a set of rules that defines that structure). The data is incomplete for a particular team member (I don't know which one, yet). If I find out it's you, I'll send you the bill. £50/hour or part thereof 
Dave Robinson, Malvern, UK
----------------------------------------[Edit 1 times, last edit by dcrobinson at Oct 12, 2016 9:30:30 AM] |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Good morning all
|
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Dave, I think that when I look at the results status on all machines including the Android there is just over 27 pages of result status, "In progress, Pending Validation, Too Late etc. Then looking at the results status by machine the first "Twoflower" shows just under 9 pages of in progress, with two or three results marked as too late, looking closely at date/time view I see that all of the results have a due date of today, so expect a lot of timed out reports. Looking at the second Twoflower there is just over 15 pages of results, some marked Valid, others showing P/Val or P/Ver and others showing In progress. Still feel that is where your discrepancy lies and am truly sorry about that,
|
||
|
|
RTS48
Veteran Cruncher Bolivia Joined: Aug 2, 2009 Post Count: 1353 Status: Offline Project Badges: 
|
Good morning all.
----------------------------------------I have been having problems with restarts wiping out progress with the CEP2 betas that have been issued recently. Seems that they checkpoint after about one and a half hours then no further checkpoints for the ensuing sixteen and a half hours. Therefore lost lots of CPU time.
Rod Peel
Santa Cruz Bolivia South America  ,  |
||
|
|
dcrobinson
Veteran Cruncher UK Joined: Mar 10, 2009 Post Count: 1176 Status: Offline Project Badges:
|
Dave, I think that when I look at the results status on all machines including the Android there is just over 27 pages of result status, "In progress, Pending Validation, Too Late etc. Then looking at the results status by machine the first "Twoflower" shows just under 9 pages of in progress, with two or three results marked as too late, looking closely at date/time view I see that all of the results have a due date of today, so expect a lot of timed out reports. Looking at the second Twoflower there is just over 15 pages of results, some marked Valid, others showing P/Val or P/Ver and others showing In progress. Still feel that is where your discrepancy lies and am truly sorry about that, My best guess is that it's because there's actually a new joiner (I'll find out later) who has joined the WCG and the Team but has not submitted any results yet. I've made some lunchtime code changes from work (the wonders of a VPN) and it's running now. All the best software has bugs in it, and I don't think it's your fault domonijo.
Dave Robinson, Malvern, UK
|
||
|
|
dcrobinson
Veteran Cruncher UK Joined: Mar 10, 2009 Post Count: 1176 Status: Offline Project Badges:
|
I've mended the stats. It appears that WCG have added some extra data to the XML that's displayed when you ask for the team member data by country. The end result was that my software, when trying to decode the data, was getting the user's registration date and then the date of the last result returned, rather than the user's statistics.
----------------------------------------I've fixed it (I had to amend the validation schema, restore the computer from a backup, and rerun everything). I've also removed the "Personal Bests" stuff from the webpage. This is because the calculation is permanently and irrevocably screwed up when you run the stats more than 12 hours late (because you get 1.5 days of data in the update instead of 1, which given that it's weekly that means about 7% inflated). I'll try to think of something else to display instead.
Dave Robinson, Malvern, UK
----------------------------------------[Edit 4 times, last edit by dcrobinson at Oct 13, 2016 6:41:41 AM] |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Good morning everyone.
It looks like my new (old machine rebuild) is now producing results are shaping up to the same rate as before the shenanigans of the last couple of weeks. Dave good to see that you have fixed the stats, WCG changing the XML schema without announcing what they were going to do. I wonder how many other team captains/stats meisters were also affected.  |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Good morning all
|
||
|
|
|