[URGENT] [IMPORTANT] ERROR: ProcessBlock() : duplicate proof-of-stake

[sigmike]

Is there anything in the recent logs?
Can you send me your log since the first mention of “ddea8c25884bede23db0” (a block you should have received before this last block you should have minted)?

[crypto_coiner]

Will get you that within 15h. Don’t have my PC with me at the moment.

[mhps]

to capture the minting address. Can anyone think of an easier way?

Choose a period when @crypto_coiner is not minting, if someone is minting with the same utxos and finds a block in this period, one should see the minting addresses @crypto_coiner has shown up in the blockchain, right?

Another thought. When I run nud with -debug -printcoinage -printcoinstake on my raspi and let the my laptop mint on the same wallet, I see in the log that the raspi finds a block (search the string “ew block found” ) and the block is orphaned a block later.

edit to add: If @crypto_coiner changes the clock of the minting computer faster by a few minutes, either the log will show the same output (with -debug -printcoinage -printcoinstake), or @crypto_coiner will start to find blocks because these blocks could be confirmed by the network before the duplicate blocks are found.

[mhps]

Seriously? Its some kind of holy grail of PoS coins and the only approaches I know of to tackle it are tendermint, DPOS or slasher (Vitalik Buterin’s consensus mechanism). All of these solutions also have their drawbacks. What is Nu’s solution?

Yeah, well the Nu/peercion devs have a solution of a problem they don’t consider really as serious as the PoWers try to make people believe it to be.
See a record of discussions of the “Nothing-at-Stake” claim and solutions to it.

[crypto_coiner]

Just sent you the recent log - the duplicate proof of stake errors are showing up, my stake is zero and I have found no block.

[crypto_coiner]

That was un-successful - “ERROR: ProcessBlock() : duplicate proof-of-stake” are back

[creon]

I’m afraid I am running out of ideas. Adding the debug flags @mhps mentioned could give some further insight. And you are absolutely sure that your system time is correct? I think windows has an option to sync it to an NTP server.

Can someone with a Windows 0.5.4 binary confirm that staking works?

[mhps]

I think windows has an option to sync it to an NTP server.

Time sync on Windows 7 on my laptop is on by default as it seems. My raspi also seems to have it. Turn it off if you do the time-drift test.

[cryptog]

Can someone with a Windows 0.5.4 binary confirm that staking works?

I m staking with no issue on win 0.5.4 client.

[Mark]

Me too.

[crypto_coiner]

I thought I would share the results of my investigation (that I shared with @sigmike in personal message) since I deem it is rather very important:

I have been taking regular back ups of my keys since sep. 2014 as a personal practice, so I have historical backups.

As an experiment, I just tried with several different backups taken in january and february, and the results are odd since the backups should be identical (I did not modify my wallet during the last 3 months, except the votes)

1/13 backup
xxxxxxx NSR
1224 tx
Last Transactions:
2015/01/13 07:39
2015/03/01

1/10 backup
xxxxxxx NSR
1224 tx
last transactions:
2015/01/10 17:58
2015/03/01

2/14 back up
xxxxxxx NSR
1254 tx
last transactions:
2015/02/14 22:24
2015/03/01

2/24 back up (the one I am using now)
xxxxxxx NSR
1258 tx
last transactions:
2015/02/21 18:29
(stops on 2/21)

Additional remarks:

• the NSR amount indicated by each backup is identical
• Stake and Unconfirmed for each backup is Null.
• I would like to mention something maybe important:
  around 2/21, when inputing a motion to vote or a custodial vote or a parking vote (i dont remember exactly) , the wallet UI allowed me to input the same vote, twice accidentally, which I removed clumsily but I felt the cancelling was not clear cut. This is obscure.

Perhaps my last key contains a duplicate vote already broadcasted which makes the network see me as broadcasting duplicate PoS, which is not the same thing at all…

I tried to mint with my 2/14 backup over the night (during 8h)
And I successfully have minted on my whole wallet.
I find new blocks and I dont have any duplicate PoS errors in my log.
My last transaction is: 2015/3/2 07:46 mint by stake 40:00nsr

Questions and enquiries:

• Hypothesis: I strongly suspect that my last backup was corrupted because of the GUI double vote bug.
• To NuShareholders: can you test whether duplicate votes (it could be on custodian, park rate, motion vote ) lead to the duplicate POS errors?
  !!BEFORE DOING THIS, BACK UP YOUR WALLETS!! This is risky and I would not be liable for any lose.

I believe I corrupted my wallet on 0.5.3 version but the bug could be still alive.

• [quote=“JordanLee, post:1, topic:1500”]
  This release contains a fix to the way motion votes are counted, resolves an obscure parking issue as well as resolving other issues. The motion vote count issue is detailed here.
  [/quote]
  What is this obscure parking issue about?

[creon]

around 2/21, when inputing a motion to vote or a custodial vote or a parking vote (i dont remember exactly) , the wallet UI allowed me to input the same vote, twice accidentally, which I removed clumsily but I felt the cancelling was not clear cut. This is obscure

If you go to Help -> Debug window -> Console and then type getvote then it will dump the voting it will write into blocks. You can check there if you still have that duplicate hash in your voting, if you don’t trust the GUI.

[crypto_coiner]

Thanks for the very useful command -
Now, I have run the command for several backup keys and it appears I dont have any duplicate vote for the 3 type of votes for any backup
Also, the backup that has stopped minting for days (froze on 2/21) indicates some blocks minted minutes ago and some stake.

Anyone has an explanation? I am puzzled…

[crypto_coiner]

Well, I give up - my last backup does not mint and I am positive about that.
I believe this is a serious bug.
I hope the open-sourcing of the code will put some light on that obscure issue.