Folding Forum

I have noticed for some time now that some of the 3 hour checkpoints report I have only completed several WU when in fact it is over a dozen. Perfect example, as of midnight 11/18 there are only 4 WU's being credited to me when in fact the Fahlogs have documented a total of 17 WU's completetd with results successfully sent. I have 9 GPU clients running 24/7 with no errors and all WU's completetd at 100%. I have taken into account the time difference between my time zone of EST and UTC time. Would somebody like to explain why I am not being credited with work that I am completing?

Double check that the client.cfg file contains the right name/team information. Best way to check is look at the start of the FAHlog.txt file.

Horvat wrote:I have noticed for some time now that some of the 3 hour checkpoints report I have only completed several WU when in fact it is over a dozen. Perfect example, as of midnight 11/18 there are only 4 WU's being credited to me when in fact the Fahlogs have documented a total of 17 WU's completetd with results successfully sent. I have 9 GPU clients running 24/7 with no errors and all WU's completetd at 100%. I have taken into account the time difference between my time zone of EST and UTC time. Would somebody like to explain why I am not being credited with work that I am completing?

Which stats site are you using?? Have you checked the offical stats at Stanford - Limk is at top of the screen

The clients are configured correctly with folding name, team number and pass key. I use extremeoc.com to track my progress and those stats have been verified.

Typically when someone has a problem of this nature, it is due to copying the FAH flies and directories from one computer to another rather than installing FAH on each machine. Each instance on different machines should have a unique User ID and each instance on the same machine must have a different Machine ID. The information is found in the log:

[16:24:48] - User ID: 4CE908AE3E8838XX
[16:24:48] - Machine ID: 3

If you have instances with the same User and Machine IDs, they will be seen as the same instance, the server will employ lost WU logic and assign them the same WU. You will get credit only for the first WU turned in. One piece of third party monitoring software, HFM will identify duplicate WUs and warn you of User/Machine ID conflicts.

The stats tend to lag the actual uploads. You do realize there are a few hours delay between when WUs are recieved and when they are credited, right? And if a particular server gets bogged down, the credits can lag many hours until it gets caught up.

How many total clients do you have, and does your user account on Stanford show that same number of active clients?

by ChasR » Thu Nov 19, 2009 12:34 pm

Each instance on different machines should have a unique User ID

Can someone clarify this for me please? I understand the machine ID but I thought the user ID was given on first run by the server. The SMP clients were assigned the same user ID's as the GPU clients on 2 of my machines. Is this normal? Should SMP/GPU clients running on the same machine have unique user ID's?

User ID should be the same on a given machine, but don't forget to set different Machine ID for each client installed on this machine.

I have checked and rechecked all the user ID's and Machine ID numbers. Those are part of the basic setup steps that I use when I setup new clients. As stated in original post I have 9 GPU clients. I also have an stanard SMP client and an SMP client running in VMware on Linux but I was not factoring those in right now because those take around 24 hours to complete. I also have to two -bigadv clients running but those are completing in 2 days 8 hours. The GPU clinets are completeing WU's every 30 minites to 3 hours depending on which WU is currently being processed.

Even taking into account the lag in registering the completed WU's, the completed WU count is still low even at the sucessive check points. I have inspected all the logs and broken them down into 3 hour sections and no matter how you slice it the logs are documenting any where from 12 to 17 completed WU's every three hours. Yet in the original scenario I gave I was only credited with 2 WU at one checkpoint and the next check point still only credited me with 17 WU's and it should have been at least over 20.

I have 4 applications running on 3 machines - one is a dual core - and monitor with Fah-Spy. I have seen it take from 24 to 36 hours to get some WUs credited, depending on the times submitted and the Stanford server loading. Lately I think the average is running about 8 to 12 hours.

It also slows down if your WU gets sent into the Collection Server instead of the original Work Server, as they still have to get moved back to the WS from the CS at a later time before getting credited to your account.

Just like a Bank Loan - your cheque's in the Mail, but isn't applied until they get it.

Well you can say what you want, something stinks at this point. All my clients are configured properly with there own ID numbers and user numbers. Also this webste shows I have 14 active clients in the last 7 days. Again how can I have 10 gpu clients running at this point, not to mention the 4 SMP clients, and only get credit for 12 work units in 6 hours when again the logs show over 24 work units completed with results sent successfully. And on top of that I still have not been credited the missing points from the other day when there was that 36 hour period of no credit for work completed at all. That should be over 100,000 points. Other team members are getting points updates and I get 0. This is starting to get ridiculous and something is seriously wrong and it is not on my end.

How many clients are you running? I see that counting the machines that have returned WUs in November, you appear to be running 24 clients. Here is a list of the most recent credit from each one. If you're missing something, please provide us with a list of the WUs that you've completed or the specific server that it came from so we know which server is having problems.

1221980A3CXXXXXXXX 4609 32 32 71 47 x86 Win 11/22/2009 6:28
1221980A3CXXXXXXXX 5771 3 15 1465 353 gpu Win 11/23/2009 2:13
210D17CF3BXXXXXXXX 5771 0 284 455 353 gpu Win 11/23/2009 2:13
210D17CF3BXXXXXXXX 5795 1 808 1 787 gpu Win 11/23/2009 2:12
210D17CF3BXXXXXXXX 5772 1 218 329 353 gpu Win 11/6/2009 14:18
2C2440AB2DXXXXXXXX 2682 6 18 2 57258.4 gpu Lin 11/22/2009 22:17
49FCB45539XXXXXXXX 2669 15 113 181 1920 x86 Lin 11/22/2009 14:14
50CC3ED575XXXXXXXX 2669 3 147 57 1920 x86 Lin 11/5/2009 0:17
51027D0D2EXXXXXXXX 2682 1 21 2 57360 gpu Lin 11/1/2009 8:42
55CB70A228XXXXXXXX 5790 16 733 2 787 gpu Win 11/11/2009 0:21
55CB70A228XXXXXXXX 5772 3 320 719 353 gpu Win 11/11/2009 2:09
55CB70A228XXXXXXXX 5790 12 521 8 787 gpu Win 11/11/2009 2:09
55CB70A228XXXXXXXX 5771 12 250 313 353 gpu Win 11/11/2009 2:09
55CB70A228XXXXXXXX 5793 2 712 19 787 gpu Win 11/11/2009 0:21
55CB70A228XXXXXXXX 5769 8 178 1695 353 gpu Win 11/11/2009 2:09
646EEABE71XXXXXXXX 5763 0 759 5 384 gpu Win 11/23/2009 2:12
646EEABE71XXXXXXXX 5771 10 159 1021 353 gpu Win 11/23/2009 0:19
646EEABE71XXXXXXXX 5770 13 85 640 353 gpu Win 11/23/2009 0:19
646EEABE71XXXXXXXX 5793 8 795 2 787 gpu Win 11/23/2009 2:12
646EEABE71XXXXXXXX 5770 6 186 379 353 gpu Win 11/13/2009 12:27
657B14F006XXXXXXXX 2683 9 10 12 58325.3 gpu Lin 11/22/2009 2:21
743271B732XXXXXXXX 5771 14 182 879 353 gpu Win 11/23/2009 0:19
743271B732XXXXXXXX 5797 18 790 5 787 gpu Win 11/23/2009 0:18

Horvat wrote:This is starting to get ridiculous and something is seriously wrong and it is not on my end.

Looks as if I got credited with two days of points yesterday, so the Team at Stanford is culling the missing and setting the contributions correctly.

Thank you, bruce for your follow-up, and thanks to the Stanford gang for the corrections. I found a pic that shows what's going on under the covers:

Horvat wrote:Well you can say what you want, something stinks at this point. All my clients are configured properly with there own ID numbers and user numbers. Also this webste shows I have 14 active clients in the last 7 days. Again how can I have 10 gpu clients running at this point, not to mention the 4 SMP clients, and only get credit for 12 work units in 6 hours when again the logs show over 24 work units completed with results sent successfully. And on top of that I still have not been credited the missing points from the other day when there was that 36 hour period of no credit for work completed at all. That should be over 100,000 points. Other team members are getting points updates and I get 0. This is starting to get ridiculous and something is seriously wrong and it is not on my end.

I'm sorry to say that I agree with Horvat. I still feel that since the credit server was fixed, there has been a rash of new missing points. I have over 25 Folding clients running as Gary480six for team 11108 and I feel like I am several thousand points short each day.
Are my clients showing rejected results or duplicated work?

I plan to re-check all my settings and user IDs over the weekend - but I do not believe that the error is on my end.

I'm looking at your EOC page, and youre resembling my page trends..

You do realize that recentely, as in around the massive update, the wus have again got varied. For the last few weeks, its been mainly, like 95%, 353 wus. But after the 22ndish, they've been sending out 353, 787, and 1888s at random.
Not sure on your cards, but my 3 gpus running mixes have "poor" days, tthe "great" days. as the 1888s take 5 to 8hrs, and throw off the nice count at each update.

I don't think anythings amiss, its just not what you're use to.