Topic: [OLD] Eligius: ASIC, no registration, no fee CPPSRB BTC + 105% PPS NMC, 877 # - page 42. (Read 458160 times)

So wk, you're saying the double payment has caused you to lose 25BTC?

Where did that extra 25BTC come from? I thought based on CPPSRB, the pool would not have any reserved fund and all shelved shares will be paid for. Its essentially max at PPS.

Since you're not honest about the problem, dont be offended if i dont trust your whole payout system. (I understand the payout method)

Might be a DNS issue.  My scripts rotate out IPs at times on our DNS. (Part of DoS protection.)  If your one machine is stuck on an old long-disabled IP try rebooting it or flushing DNS.

 

Weird, I switched to GETWORK and it's working now. I tried stratum and still no luck on two out of three computers. I have been using Eligius since December 10 without any connection issues.

Nothing going on that should have caused any connectivity issues on this end.

I just lost connection to eligius stratum on two out of three computers at home.

http://eligius.st/~wizkid057/newstats/userstats.php/15aJAPshhWsNJRYx4FnjYPd9fer8XM1cNT

One keeps hashing like nothing happened. When I tried to restart bfgminer on the two failed ones I get "Pool Setup failed". I switched to GHash.IO for the time being.

Thanks for the quick reply!  I'll point my miners back at Eligius.  I know my .2 BTC that I'm working off is hardly what you're worried about, but every little bit helps. 

No need.

Guess the cat is out of the bag sooner than I had hoped.  Anyway... Abridged version:  In the confusion surrounding last night's emergency I made a mistake that allowed the reward system to overpay some miners.  
No miners have lost any income due to this and it is business as usual.

Details: The database issue happened right around the time we mined block #277343.  Since the CPPSRB code couldn't read the database at the time, it wasn't able to count the payouts in that block's coinbase transaction.  CPPSRB went into fail-safe mode pending my intervention.  

I was automatically notified very soon after of the database issue.  I saved state and shutdown CPPSRB while it was in fail-safe mode without paying attention to the fact that I hadn't allowed it to tally this one block's earnings yet.

Then I proceeded to correct the hardware issue, etc, etc........ and I completely forgot CPPSRB was in fail-safe when I saved state and shut it down.  Without thinking when I started it back up, I ignored its cries about fail-safe and allowed it to load its state and continue...... without it tallying the fail-safe'd block.

I actually realized shortly after that I had made this error.  No harm done yet.  I went to shutdown CPPSRB to correct it and the moment I did we found block 277361..... containing a set of payouts remarkably similar to those in block 277343.  Some people had then officially been double paid.

After the last issue I had with a double payment months and months ago (resulting in a final roughly 2 BTC loss that I covered at the time), I had coded in some handling of negative balances, allowing miners to "work off" an overpayment.  However, I didn't think about how the stats code would react, since they assume a balance can never be negative.  This seems to manifest on the stats as a messed up maximum reward graph and never having any balance stick until the over-payment is covered.

Unfortunately, the Bitcoin community has not been known for its honest users, overall.  So, my apologies in having to withhold this information, and I hope the reasoning is understood.  I'm sure if I had immediately posted something about this that most people who were double paid would have took off with the over-payment never to return.  It is sad, but true.  Fortunately the majority of affected miners continued to mine as usual and a good portion of this error has corrected itself.

In any case, I know there are going to be a bunch of people who now will do exactly as I initially feared and just flee with whatever part of their over-payment they have, but I kindly ask that anyone affected please have some respect, realize that I am a fallible human being that makes mistakes, consider any mistaken payment as a 0% interest payday loan, and simply continue to mine as normal so that the overage is covered because any remaining overage will end up coming out of donated funds as I don't have a huge pile of income from pool fees to fall back on for something like this.

I sincerely apologize for the issue and confusion caused by it.  Regardless of how much of these funds are unrecoverable, I won't allow any miners to be adversely affected by this mistake and will ensure that it is covered.  (As of this writing about 80% of the over-payment has been repaid through mining so far... I expect that to increase further when more shelved shares are paid.)

I have also put in code that will yell at me in the future if I try to dump CPPSRB's state while it is in fail-safe mode to prevent this particular oversight from ever happening again.


Long story short.... business as usual.  Pool is up and running, shares are tallied, blocks are being found, everything is working as expected.  For the few miners affected by this, please accept my apologies for the confusion and I hope you will do the right thing and continue to mine.

Thanks for mining.

-wk

Edit: Clarification:  Miners/share log shares were credited for block 277343 as normal.  Also, I will correct the stats to show block 277343 soon.

Guess I'll move my miners over to Bitminter until this is sorted out.  Hope I didn't lose 1/4 of a day here.  

OK, wizkid responded to my question, he's checking it, but no shares are lost, everything will be calculated properly for correct payouts.
no worries...

Oddly the last payout is actually exact my estimated total just b4 the server went down. I still kept that page open until the server was fixed (thanks to the lack of stats regarding shares). So i dont see how thats double payment.

I have two payments that are basically the same with-in two hours of each other.  My "Estimated Change" has been the same for at least 8 hours.



P.S.  Ignore the drop in my hash rate, I just rebooted some of my miners.  

I'm really tired of these lame stats. No shares per round, no shares per day or hr, and no total shelved shares. Its hard to keep track of my submitted shares.

From what i read, my shelved shares will always be paid out. Yet for the past 2 months my shares reward has never been above 99%. Either the pool was never "lucky" or my shelved shares were never paid fully.

As for the NMC, its completely in the dark. Just paid out, noway to know if its counted correctly.

That sucks


Everyone should stop mining until this is solved.

Double payment? you meant you got paid twice of what you earned?

I dont see any double payment. The last payment i got was what i earned b4 the server was down. I switched to my backup pool and didnt switch back until 9hrs ago.

Is your unpaid balance still zero? but you dont see any payout?

same here...

Same here.  It looks like I got a double payment though (I was in the queue when the pool died last night).  So I'm guessing that maybe the system is "catching up" to get me back in line where I should be?

Is there something wrong with the reward?

Mine keep resetting. The "as of last block" is still at zero and i dont see those payouts.

The balance graph is also messed up ... it doesnt show any BTC paidout or earned.....

Hashrate reports fine.

 

So after checking further, i have mined for the past 8hrs without any reward.

The last paid out was after the server got back up online. I switched back from my backup pool 8hrs ago and my unpaid balance is still zero.

Can you explain pls?


EDIT: i'm not even in the payout queue even tho the stats says i would be in the next block.

thanks wizkid057
i am using last version but continues issues

Currently working on a rewrite of the cgminer driver for KnC, but my time is limited the next few days, so, I don't expect to complete it until mid next week.  There may still be an issue with the KnC FPGA code, however, but I won't know for sure until i get my standalone test code done.

Last I checked KnC pushed a commit to their cgminer branch that fixed some of the HW errors, but does not fix the flush issue.

-wk

any fix for errors kncminer?