An update has been made to the Payout History page to make it easier to use for long time miners.  The page will now only list your 10 most recent payouts for BTC and 10 for NMC.  A full history log can be accessed clicking the link under each section.  This will load a page of comma separated values, which also includes the Transaction ID for the payment.

This can be especially useful for anybody that has had issues with a coin exchange not recognizing all of their deposits, since a Transaction ID can provide them with proof of the time, amount, wallet, and confirmations of any given deposit.

A few optimizations were done to improve load time on the Earnings History page.  After so many difficulty switches, the queries that were originally in place were starting to take longer than they should have.

Using this update [thread has been quiet lately] to remind EU users that you can access a pool server in Germany using de.btcguild.com .  This may even be preferable to any users that occasionally have connection issues between themselves and the Chicago based servers.

The daily volume on the exchanges has been significantly higher than the coins mined per day for quite some time.  Cutting the mining rate in half is not cutting the TRADING SUPPLY in half.  It will have very little impact.

BTC Guild luck has been pretty abysmal the last two days, but I'd still not call it fishy.  Every pool has good days and bad days, we've just happened to have a few pools get their bad luck at the same time.

Username would be exactly what is shown on the manage workers screen, like:

eleuthria_mediacenter

+1.  With the amount of time wasted and stress suffered from botnets, I feel like I've aged a good 10 years since last June.

You have been promptly ignored .   Glad to hear you sorted it out.

With the network growing again after the large slump last fall, I have enabled a server for miners in Europe, located in Germany.  Some miners are still using the old address and have already started using it [sorry to those miners who had some connection issues earlier while I was changing the DNS].

If you're in Europe or Asia, you should try the new server out:   de.btcguild.com  (port 8332 as usual).

This thread really needs to be more visible:  Perhaps a giant bold red link at the top of the forum?

BTC Guild will be pushing the patch tomorrow, and my understanding is both Slush & Deepbit will be shortly as well, which means the chain will have majority support.  That doesn't solve the problem of the other 30-40% of bitcoin miners that may not even know about the issue due to lack of visibility.

Bumping up!  New IP block assigned, and ready to keep growing.

A little different, my understanding is Luke-Jr was actually ignoring other nodes blocks and just kept building upon his own?

Either way, the suppression attack is what I meant by "strangulation attack" - It slows down the networks ability to process transactions because a portion of blocks ignore all TXes, which also raises the difficulty.  This means that transactions are now slowed down by *AT LEAST* their portion of the network due to legit miners now running at a higher difficulty.  It could slow down the network even more if enough nodes were doing this type of "attack" and created a backlog of transactions where legit nodes could not wrap them up into blocks as fast as they're coming in.

Pool down temporarily.  Major problem caused by bitcoin v6 and BIP16 being enforced before its supposed to.


EDIT:  Pool back online.  One of the nodes had not restarted with '-paytoscripthashtime=1333238400', causing it to be working on a hard chain fork.

Just to be completely clear, everything I'm posting is pure theory.  There is nothing that says 88.6.216.9 is a botnet bitcoind node, or has any setup like what I described above.  A lot of what I've posted came from originally looking at what it would take to redesign the current miner/pool communication model, but modified to how it could work if designed for a malicious user.

I suppose in the end this is preferred to the alternative:  1 million bot miners bouncing between pools and taking them offline with the obscene levels of overhead/traffic it would generate if acting legit.

Tough to say exactly what you'd need to respond to it, but just some rough numbers.  Assuming a 1 million bot botnet (which might be conservative):

If each bot checks in once per minute to see if the block has changed, you're looking at 16.6k connections per second.  Lets assume its still using HTTP just for simplicity.  The response will be ~200 bytes including headers/overhead give or take if no block change has occurred, and ~300 bytes if it did (a response would have to include wallet address, prevblockhash, and difficulty).  A block change every 10 minutes would mean every 10 minutes you have 2375 bytes of data being transferred, per bot.

That comes out to 1.95 gigabytes per 10 minutes, or 27 mbps.  You'd probably have a fairly even distribution since each bot started at different times, though it would actually be 25 mbps 9 minutes out of 10, and 38 mbps for the minute after a block change.

16.6k connections per second is the real bottleneck there.  However, if the system was distributed to where the slaves contact any one of 25 cheap $5 VPSes, you have about 600 connections per second, and each VPS would only need ~1 mbps of transfer.  Basically 1 hour worth of bot activity would likely pay for an entire month of servers.  They could add/remove servers quickly and use DNS as a method for distributing the bots without a central control location.


The above is why I stated this type of activity is scary, even though it is inevitable.  You're looking at a system that could easily grow to make up a majority of the network, for a few hundred dollars a month in expenses to the operator.  It can, and eventually will make mining unprofitable for any legitimate party*.  It would be good for network security, unless it was turned into a 51% attack attempt.  Alternatively, it could be a sort of strangulation attack, where the the botnet is pumping out 0 tx blocks, and difficulty gets driven up to the point legit blocks can't fit all the transactions into them, causing a growing backlog.

*: By "any legitimate party" I'm not including very small miners. While some people do have free electricity, very few can grow beyond a few GPUs before a landlord gets suspicious/angry.  And even then, it could reach the point where the mining income doesn't match the cost to replace dying hardware.

I stated previously the kind of hardware it would take for a botnet to proxy around and use a traditional pool.  THIS type of behavior would be drastically different.  A custom coded server that responds to a bot with:  UniqueWallet,PrevBlockHash,Difficulty, and a solo miner that takes that information and continually generates its own work instead of relying on a central bitcoind to generate work on its behalf.

Very interested to find out what this IP is.  Been looking at it, and the way the transactions work, paired with the size makes me fear it's a botnet.  The easiest way to run a botnet without a pool:  Make a special server that distributes a wallet address, a hash of the previous block, and the current difficulty, and let the bot do all its own work instead of the grabbing work from bitcoind.

If it's not a botnet, then we've got a major player coming in that is effectively "attacking" {generating blocks but purposely not putting in transactions} the network.  Obviously luck could be a factor, but the current generation rate is well over 1TH/sec.

It actually disappeared briefly when switching from Proportional to PPS.  But that only lasted a week I believe.

I've had this as the IRC message for a while, but realized I have not posted it in the forums yet:

I have been running some special reports every few days to help detect malicious users [likely botnets].  When a user pops up that has many of the signs of a botnet, they are banned.  This has had a fairly significant impact on the pool performance in the last two weeks, as the amount of work/connections being run through the pool have decreased by roughly 40%, even though speed has been fairly constant.

For a user banned in this way, it's as if their account does not exist when they try to log in.  The number of accounts banned this way has been decreasing with each subsequent wave, which is good since it looks like those types of people are leaving the servers and not coming back.  If you ever have a problem logging in, please contact me with your username and I will look into it/remove the ban.  All of your data is still there, just inaccessible until the ban is removed.

The process has been manual so far, but it is very close to becoming automated.  I've been waiting to see if any false positives show up, but so far not a single legitimate user has come forward as a result of a banned account.

Problem seems to be fixed for Smoov.  Everybody using GUIMiner should update to the latest version (released 2/19/12).  Link to the thread here: https://bitcointalk.org/index.php?topic=3878.0

The versions over the last few months have had a bug with the poclbm miner (default miner of GUIMiner).