Duplicate WUs?

P

plext0r

[H]ard DCOTM x3
Joined
Dec 1, 2009
Messages
780
Yesterday I was looking at my HFM.net monitor and I had four machines with duplicate WUs (two each with two different WUs). In the past, I've seen this once or twice and just left things alone and everything worked out (except for the possible missed points).

Is there something I should search for in the logs to narrow down why two different boxes are getting the same WU at different times of the day?

Thanks!

[URL="
 
Last edited:
Normal only one (1) copy of a WU exists but if somebody dumps a WU (inc crashes) or lets it timeout (not final deadline) then it is cloned a few times (5-10) and sent off, if any one of them fails to come back same thing happens.

That is why dumping a WU is so bad, it means minimum 5 other people have to spend time doing it when they could be doing other WUs.
 
I would check your logs on those 2 wus if you can. I looked up both in the wu db and only 1 person for each is shown as submitting them. In both cases it isn't you. Something isn't right.

Your WU (P8101 R4 C0 G32) was added to the stats database on 2012-03-26 11:08:37 for 343694 points of credit.
Your WU (P8569 R0 C3 G15) was added to the stats database on 2013-05-14 06:11:11 for 20192.2 points of credit.

Edit: If you expand the column with the wu info it wont chop off the Gen number. I'm not sure if that 2nd wu was actually Gen15.
 
Last edited:
Identical machinedependent.dat, rather?
 
I think to receive duplicate WU, it has to be identical machineid + machinedpendent.dat.
 
While those might not be the correct WU numbers, that orange highlighting is (I think) HFM saying something is wrong. So they may be duplicate WUs. How did you install these clients?

Edit: Where these WU running at the same time (or did they at least overlap)?
 
tear said:
Identical machinedependent.dat, rather?
Click to expand...
Yes indeed. V6 stores User ID as machinedependent.dat on Linux. (Windows uses registry)

kevinksu said:
I think to receive duplicate WU, it has to be identical machineid + machinedependent.dat.
Click to expand...
Correct. Stanford identifies each client by it's CPUID, a combination of User ID & Machine ID.

EXT64 said:
While those might not be the correct WU numbers, that orange highlighting is (I think) HFM saying something is wrong. So they may be duplicate WUs.
Click to expand...
Back when I was running several dozen GPU clients I would get duplicate WUs
occasionally, and HFM colored them the same why. Haven't had any duplicates
running BA, but I'm only running 3 BA folders so my odds are much lower.
 
You must log in or register to reply here.
Back
Top