This.


I know that, unfortunately it's pretty complex stuff to explain (exponential functio, reason for making older shares less valuable, renormalization) and I failed with some easy description for 'common users'. But because expected payout is +- the same as for share based calculation (without pool hoping), I feel like shares/total shares * 49 is correct enough for average user.

I cross-checked everything on pool and there's nothing suspicious. It's just long round :-). It is currently around 6x more than difficulty (should be 1x in average), but it's still not the worst round in history.

Thanks, I'll give a try!

Fine, I set up repository on http://www.assembla.com/code/sierrachartfeed/mercurial/nodes

Yes, it is targeted to end users on Windows - they probably don't want to install any dependencies, just download & run.


Yes, read top post, sources are available. However, you have to install python for using this.

Thats higher difficulty & bad luck combined.

Block looks fine, there wasn't any pool issue during this time. Probably network issue on your side. Do you have any log of network activity?

Yes, bridge is using websocket for mtgox and bitcoincharts polling for the rest. This error is definitely related to some network troubles, but I should make better exception handling to not crash whole bridge when it happen.

I didn't read all threads about mtgox claims - but is there any known reason why only one account per email can be processed? Mtgox confirmed my one of three registered accounts, but there are no funds; and two other with significant money on them are blocked. This is... simply weird. I'm pretty patient person, but keeping my money blocked even when other will be happily trading makes me uncomfortable.

Hm, unstable connection?

Why we need 'lock'? I don't see any advantage in that. Email confirmation isn't enough? Afaik it works very well for pool users.

Some users has changed wallet, because attacker entered their account before I released email confirmation. But I detected many intrusions and cancelled attacker address on those accounts. But of course everybody have to check his account if wallet is correct.

As far as I can say, pool accounts are now safe, even if attacker know login/password for them.

I found that at least one attacker compromised over 100 accounts, which is insane; it means that so many people had same login/password for both sites (pool, mtgox) and they even didn't changed password after strong recommendation in mtgox mass email.

I removed this wallet (http://blockexplorer.com/address/13F4yQfbzA6h2xiyqKVmhGr95zo5DGkK5R) from all pool accounts and thanks to email confirmations of wallet change, those attacks are over. Fortunately only 4.39 BTC was lost in total...

If you find that your wallet on pool profile is "Enter your address here", it means that your account was compromised and I removed attacker's wallet from your profile.

I just released email confirmations for wallet changes on profile page.

About blocks #5898 and #5896; They were very short, but both are invalid.  Probability of finding two solutions by two pool backends at same time (and twice in the row) is extremely small, but as you see, it may happen.

Pool backends are independent machines, which synchronize data as soon as possible, but broadcasting bitcoin block information is pretty complex stuff, so it may take one or two seconds. In those two blocks, two miners found solution too fast in a row and backends didn't notice that there is already valid solution for this block, so they decide to accept this solution, too. This leads to two 'invalid' blocks, but no (or only very small fraction) of mining power was wasted. I know that two 'invalid blocks' in a row looks weird, but there's not a simple solution to fix this issue and reject invalid block solution when another machine just found valid one.

It's well known bug, but only cosmetic one. Link to blockexplorer is correct.

It takes some time to calculate everything, reward of last found block can change few times.

Hm, really crazy. Great that you fixed it.

Weird. Do you see any error? Can you ping api.bitcoin.cz? (IP should be ****.32)

Yes . One backend was blocked, so part of workers didn't worked.

Again - did somebody changed your worker passwords? Or it was just a temporary error which you saw in miner?

I don't think somebody entered database.

slush

One backend had troubles and monitoring didn't catch it. It's repaired now. I'm sorry for troubles...