Topic: [ANN][YAC] YACoin ongoing development - page 148. (Read 380091 times)

This may be partially correct, but I'm not in a good spot right now to go code diving.  The issue may be that if someone stakes PoS blocks while off-network, they may be able to trigger a reorg and orphan PoW blocks if they achieve a higher cumulative work for their off-network blockchain (meaning that the rest of the network did not stake enough PoS blocks in the meantime).  Someone may have just lucked out in this case and managed to stake two PoS blocks while the rest of the network had only staked one.  Not sure yet.  Someone who can go code diving should check how PoS blocks affect the cumulative work of a chain and under what circumstances someone off-network can make their own chain with higher cumulative work without actually having 51% of the hash power (or 51% of the coin stakes being generated).

From my recent investigation in the Elacoin client source while Elacoin was being 51%'d over and over, the answers will probably be in main.cpp in the AcceptBlock(), ConnectBlock() and CheckBlock() functions, or wherever else the Reorganize() function gets called from.  Normal behavior in Bitcoin and Litecoin derived coins is that the chain with the most cumulative work (difficulty) will win a reorg and block height has nothing to do with it, but I have not examined that part of the code on any PoS coin to see how PoS blocks factor into which chain has the highest cumulative work.

Um guys.... this isn't good


They are literally taking all the blocks. I mean like nearly all of them. I have been watching them for a while and I continue to see almost all the their block confirmation values are consecutive numbers.  

Seems every proof of stake will make proof of work invalid.

Your results in the client are going to be different depending on which client you have.  Because of the low CPU hashrate as N went up, WindMaster set it to report an average, but the hashrate is still pretty swingy.  That's why I was using the stats on yacexplorer.tk

in my client it was 72MH 5 hours ago and didnt drop untill the last 2 or more hours to 36MH and now its back up so...

sairon: In yacexplorer, how often is the top stats (richest addresses, etc) updated? Looks like the numbers up there are form 6-1. Any way to update more frequently?

In yacexplorer.tk graph, there's an extreme spike in recent hash rate, what's wrong with that?

well, if there's only 36MH on the network, then there's the problem, yac.ltcoin.net is currently reporting is has 37 MH/sec...  Actually, this could be problematic as well as the mean hashrate is 72 MH/sec as reported on yacexplorer.tk

Is this may be the reason for my problem?
This transaction is still not confirmed, and my wallet always says to stake=0 after -rescan and -reindex. And impossible to find this transaction in yacexplorer.


https://bitcointalksearch.org/topic/m.2461662

Thank you!!

Thanks YacLives and Bitcoin Megastore for your debug logs.

PoS blocks 91036 and 91037 were staked by someone operating off-network (whether intentionally or unintentionally) and then reintroduced to the network, triggering the reorg and orphaning 75 blocks.  Anyone here have really good familiarity with the Novacoin PoS code?

for those can have a look at "what have happened in an hour"

http://yac.ltcoin.net/blocks.php

So far it looks like everyone is actually on the same block # and blockchain that I'm on with the newest code, so most likely something else occurred unrelated to the code changes.

Anyone have a copy of their debug log going back a couple hours so we can get a better look at what happened?  I see eule posted the reorg itself but I'd like to get a look at what came before it.  I don't have a debug log handy since I wasn't running yacoind continuously while I was preparing the commits to Github.

Just noting that my (old) client and yacexplorer are now at the same number of total blocks (91121).

FYI yac.coinmine.pl has 19mh of the 36 on the network as of right now

my pool use original source code when it announce

is this a problem?

Where does one vote for best alt dev?

I'm using the original client too

Roger that.  Looks like the timing was a coincidence then, and I won't revert the scrypt-jane library changes to the Github repository.

On that note, are there any win64 binaries available anywhere for these new client builds?

i have a lower block count and im on the alpha client not the new beta version


and i think its unrelated to your update but YAC it has been very strange in the last 16 hours