[1500 TH] p2pool: Decentralized, DoS-resistant, Hop-Proof pool

[nreal]

I got error when I found a share today:

Code:

2013-12-12 01:01:36.423805 P2Pool: 17345 shares in chain (17349 verified/17349 total) Peers: 31 (21 incoming)
2013-12-12 01:01:36.423930  Local: 2356GH/s in last 10.0 minutes Local dead on arrival: ~1.4% (0-3%) Expected time to share: 29.2 minutes
2013-12-12 01:01:36.423978  Shares: 0 (0 orphan, 0 dead) Stale rate: ??? Efficiency: ??? Current payout: 0.0000 BTC
2013-12-12 01:01:36.424015  Pool: 116TH/s Stale rate: 20.5% Expected time to block: 9.3 hours
2013-12-12 01:01:39.436918 P2Pool: 17345 shares in chain (17349 verified/17349 total) Peers: 31 (21 incoming)
2013-12-12 01:01:39.437093  Local: 2357GH/s in last 10.0 minutes Local dead on arrival: ~1.3% (0-3%) Expected time to share: 29.2 minutes
2013-12-12 01:01:39.437174  Shares: 0 (0 orphan, 0 dead) Stale rate: ??? Efficiency: ??? Current payout: 0.0000 BTC
2013-12-12 01:01:39.437264  Pool: 116TH/s Stale rate: 20.5% Expected time to block: 9.3 hours
2013-12-12 01:01:41.593547 GOT SHARE! Jupiter 23877a5a prev a9d32777 age 80.86s
2013-12-12 01:01:41.622728 > Error while processing Event callbacks:
2013-12-12 01:01:41.622811 > Traceback (most recent call last):
2013-12-12 01:01:41.622846 >   File "/home/user/p2pool/p2pool/bitcoin/worker_interface.py", line 136, in <lambda>
2013-12-12 01:01:41.622882 >     lambda header, user, coinbase_nonce: handler(header, user, pack.IntType(self._my_bits).pack(nonce) + coinbase_nonce),
2013-12-12 01:01:41.622915 >   File "/home/user/p2pool/p2pool/work.py", line 403, in got_response
2013-12-12 01:01:41.622981 >     self.node.set_best_share()
2013-12-12 01:01:41.623012 >   File "/home/user/p2pool/p2pool/node.py", line 297, in set_best_share
2013-12-12 01:01:41.623043 >     self.best_share_var.set(best)
2013-12-12 01:01:41.623073 >   File "/home/user/p2pool/p2pool/util/variable.py", line 74, in set
2013-12-12 01:01:41.623103 >     self.changed.happened(value)
2013-12-12 01:01:41.623133 > --- <exception caught here> ---
2013-12-12 01:01:41.623162 >   File "/home/user/p2pool/p2pool/util/variable.py", line 42, in happened
2013-12-12 01:01:41.623192 >     func(*event)
2013-12-12 01:01:41.623232 >   File "/home/user/p2pool/p2pool/node.py", line 96, in broadcast_share
2013-12-12 01:01:41.623271 >     peer.sendShares([share for share in shares if share.peer_addr != peer.addr], self.node.tracker, self.node.known_txs_var.valu
e, include_txs_with=[share_hash])
2013-12-12 01:01:41.623322 >   File "/home/user/p2pool/p2pool/p2p.py", line 291, in sendShares
2013-12-12 01:01:41.623355 >     assert tx_hash in known_txs, 'tried to broadcast share without knowing all its new transactions'
2013-12-12 01:01:41.623394 > exceptions.AssertionError: tried to broadcast share without knowing all its new transactions
2013-12-12 01:01:41.660075 New work for worker! Difficulty: 599.266435 Share difficulty: 1076527.143944 Total block value: 25.061445 BTC including 243 transa
ctions
2013-12-12 01:01:41.691285 New work for worker! Difficulty: 200.000000 Share difficulty: 1076527.143944 Total block value: 25.061445 BTC including 243 transa
ctions
2013-12-12 01:01:42.448631 P2Pool: 17346 shares in chain (17350 verified/17350 total) Peers: 31 (21 incoming)
2013-12-12 01:01:42.448763  Local: 2360GH/s in last 10.0 minutes Local dead on arrival: ~1.5% (1-3%) Expected time to share: 29.2 minutes
2013-12-12 01:01:42.448800  Shares: 1 (0 orphan, 0 dead) Stale rate: ~0.0% (0-80%) Efficiency: ~125.8% (25-126%) Current payout: 0.0080 BTC
2013-12-12 01:01:42.448837  Pool: 116TH/s Stale rate: 20.5% Expected time to block: 9.3 hours
2013-12-12 01:01:45.462377 P2Pool: 17346 shares in chain (17350 verified/17350 total) Peers: 31 (21 incoming)
2013-12-12 01:01:45.462555  Local: 2369GH/s in last 10.0 minutes Local dead on arrival: ~1.5% (1-3%) Expected time to share: 29.2 minutes
2013-12-12 01:01:45.462638  Shares: 1 (0 orphan, 0 dead) Stale rate: ~0.0% (0-80%) Efficiency: ~125.8% (25-126%) Current payout: 0.0080 BTC

Then found another share and it was fine:

Code:

2013-12-12 01:04:22.130684  Local: 2329GH/s in last 10.0 minutes Local dead on arrival: ~1.6% (1-3%) Expected time to share: 30.1 minutes
2013-12-12 01:04:22.130717  Shares: 1 (1 orphan, 0 dead) Stale rate: ~100.0% (20-100%) Efficiency: ~0.0% (0-100%) Current payout: 0.0000 BTC
2013-12-12 01:04:22.130754  Pool: 115TH/s Stale rate: 20.5% Expected time to block: 9.4 hours
2013-12-12 01:04:22.335429 New work for worker! Difficulty: 691.171186 Share difficulty: 1241624.893775 Total block value: 25.087831 BTC including 352 transa
ctions
2013-12-12 01:04:22.377099 New work for worker! Difficulty: 200.000000 Share difficulty: 1241624.893775 Total block value: 25.087831 BTC including 352 transa
ctions
2013-12-12 01:04:25.143566 P2Pool: 17351 shares in chain (17355 verified/17355 total) Peers: 31 (21 incoming)
2013-12-12 01:04:25.143681  Local: 2325GH/s in last 10.0 minutes Local dead on arrival: ~1.7% (1-3%) Expected time to share: 30.2 minutes
2013-12-12 01:04:25.143714  Shares: 1 (1 orphan, 0 dead) Stale rate: ~100.0% (20-100%) Efficiency: ~0.0% (0-100%) Current payout: 0.0000 BTC
2013-12-12 01:04:25.143750  Pool: 120TH/s Stale rate: 20.5% Expected time to block: 9.0 hours
2013-12-12 01:04:27.234256 GOT SHARE! Jupiter3 89d9ea47 prev faed7137 age 4.87s
2013-12-12 01:04:27.349609 New work for worker! Difficulty: 614.131984 Share difficulty: 1103231.495636 Total block value: 25.087831 BTC including 352 transa
ctions
2013-12-12 01:04:27.391143 New work for worker! Difficulty: 200.000000 Share difficulty: 1103231.495636 Total block value: 25.087831 BTC including 352 transa
ctions
2013-12-12 01:04:28.156049 P2Pool: 17352 shares in chain (17356 verified/17356 total) Peers: 31 (21 incoming)
2013-12-12 01:04:28.156144  Local: 2327GH/s in last 10.0 minutes Local dead on arrival: ~1.7% (1-3%) Expected time to share: 30.2 minutes
2013-12-12 01:04:28.156176  Shares: 2 (1 orphan, 0 dead) Stale rate: ~50.0% (9-91%) Efficiency: ~63.3% (11-115%) Current payout: 0.0114 BTC
2013-12-12 01:04:28.156213  Pool: 121TH/s Stale rate: 21.1% Expected time to block: 8.9 hours
2013-12-12 01:04:30.834454 Peer sent entire transaction 90e0bb504ac630d2e08ddf489731341236812c22179e74fedde6a6b5d96282ad that was already received
2013-12-12 01:04:31.166570 P2Pool: 17352 shares in chain (17356 verified/17356 total) Peers: 31 (21 incoming)

Also, adjusting share diff works for my node, just fine. p2pool gives work to my workers and reporting share time:

Code:

2013-12-12 01:06:00.703104 Got new merged mining work!
2013-12-12 01:06:00.740053 New work for worker! Difficulty: 570.368318 Share difficulty: 1024614.009517 Total block value: 25.021775 BTC including 157 transactions
2013-12-12 01:06:00.767854 New work for worker! Difficulty: 200.000000 Share difficulty: 1024614.009517 Total block value: 25.021775 BTC including 157 transactions
2013-12-12 01:06:01.621688 P2Pool: 17351 shares in chain (17355 verified/17355 total) Peers: 31 (21 incoming)
2013-12-12 01:06:01.621786  Local: 2352GH/s in last 10.0 minutes Local dead on arrival: ~1.8% (1-3%) Expected time to share: 30.4 minutes

But on webpage it says:

Code:

Pool rate: 121TH/s (21% DOA+orphan) Share difficulty: 487000

Node uptime: 4.0 days Peers: 10 out, 21 in

Local rate: 2.41TH/s (1.5% DOA) Expected time to share: 14.5 minutes

So we have different share diff and ETA for workers in p2pool log, and different share diff nad ETA on p2pool front page. But adjusting itself works very good, I have Expected time to share in p2pool log always auto-adjusted to ~30 minutes.

EDIT:
Both shares I found has been orphaned couple of minutes later!!! :O With series of errors. Whole log here:
http://lenny.dnsd.me/p2pool.log

Shares: 2 (2 orphan, 0 dead) Stale rate: ~100.0% (34-100%) Efficiency: ~0.0% (0-80%) Current payout: 0.0000 BTC
  Can someone explain how it happened?

I did have same thing going with my node, what i did was rebuild everything needed with p2pool with useflag soap enabled, havent seen that error since. Not sure if it was that, but that error doesnt show up anymore..

[nreal]

To start p2pool as daemon, not sure if its explained somewhere but didnt found it.

One can start it like this: nohup python run_p2pool.py ....... > /dev/null 2>&1 &

So then you can restart X or even logout without killing p2pool.

> /dev/null 2>&1 & at the end makes nohup not to log anything and the & makes it run as a daemon

If you want to read that useless log it can be done  - tail -f /p2pool/data/bitcoin/log

[domob]

This has probably been asked before somewhere, but please excuse that I'm not going to read through all pages here.  It seems to me that you need a large hash rate in order to use P2Pool reasonably, right?  The current rate of P2Pool is reported as 121 TH/s.  It is scheduled to produce 8,640 shares per day (?), which means that on average, to get a share per day, you need 14 GH/s.  And one share per day on average is not that great, IMHO, in order to get reasonably steady payouts.

Am I correct to assume that it makes no sense to use P2Pool without 10-100 GH/s?  I would love to use it instead of a centralised pool, but it really seems not a good idea for me (I have around 1.5 GH/s, not to make a profit, more for the fun - but I still get a couple mBTC per week, which is not bad).  Wouldn't it be possible in theory (if the system is changed a bit, I understand that) to create a P2Pool for P2Pool and so on, building a hierarchy of networks, which allow even smaller rates?  It probably would not be worth the effort just to make it usable to small miners, I guess.

[K1773R]

This has probably been asked before somewhere, but please excuse that I'm not going to read through all pages here.  It seems to me that you need a large hash rate in order to use P2Pool reasonably, right?  The current rate of P2Pool is reported as 121 TH/s.  It is scheduled to produce 8,640 shares per day (?), which means that on average, to get a share per day, you need 14 GH/s.  And one share per day on average is not that great, IMHO, in order to get reasonably steady payouts.

Am I correct to assume that it makes no sense to use P2Pool without 10-100 GH/s?  I would love to use it instead of a centralised pool, but it really seems not a good idea for me (I have around 1.5 GH/s, not to make a profit, more for the fun - but I still get a couple mBTC per week, which is not bad).  Wouldn't it be possible in theory (if the system is changed a bit, I understand that) to create a P2Pool for P2Pool and so on, building a hierarchy of networks, which allow even smaller rates?  It probably would not be worth the effort just to make it usable to small miners, I guess.

p2pool isnt good at payout out steady in this case, but you get more coins than anywhere long-term.

[TierNolan]

Wouldn't it be possible in theory (if the system is changed a bit, I understand that) to create a P2Pool for P2Pool and so on, building a hierarchy of networks, which allow even smaller rates?  It probably would not be worth the effort just to make it usable to small miners, I guess.

The problem with hierarchical p2pool is that the system assumes you can pay to multiple addresses from the coin base.

When you win a share with p2pool, you tell the system what address you want to pay to.

When anyone finds a share that is good enough for bitcoin itself, then that blocks pays out to the most recent N p2pool shareholders.

This tends to lead to lots of small payouts.

To run a 2nd level p2pool, you would do the same.

blocks -> bitcoin main chain blocks
shares -> p2pool shares
share-lites -> 2nd level shares

The share-lites would be a 2nd level of p2pool.  If someone produces a share-lite that meets the actual p2pool share difficulty, then they are supposed to share that share with the last N share-lite share holders.

The problem is that p2pool doesn't let you specify a list of payout addresses.

If it did, then there would be a very large number of shareholders when a block is finally found.  Many miners would end up with dust payments.

The way centralised pools do it is that they don't pay out to all miners for every block.  They combine lots of your (small) payouts into a single payment.  You could say that you want payments only if you are owed at least 0.01BTC.

P2pool has no central authority, so it can't store payments like that, in order to combine them.

It could do some smoothing by storing some debt in the sharechain itself.  If you are skipped from payment, you are paid when the next block arrives.  Getting that algorithm right is hard.  If the pool stops working, then some people will end up not being paid.  Maybe donors could donate money to the sharechain, so it has a pool of cash.

[domob]

It could do some smoothing by storing some debt in the sharechain itself.  If you are skipped from payment, you are paid when the next block arrives.  Getting that algorithm right is hard.  If the pool stops working, then some people will end up not being paid.  Maybe donors could donate money to the sharechain, so it has a pool of cash.

Yes, I think that's basically what would be required:  Some kind of "P2Pool coin", where you can build up a balance in P2Pool until it is worth to issue a Bitcoin payment.  Without thinking about it further it seems as if this should be possible technically to get right in a decentralised manner somehow, but I understand that this is probably way too much effort and complication for its benefit.

[TierNolan]

Yes, I think that's basically what would be required:  Some kind of "P2Pool coin", where you can build up a balance in P2Pool until it is worth to issue a Bitcoin payment.  Without thinking about it further it seems as if this should be possible technically to get right in a decentralised manner somehow, but I understand that this is probably way too much effort and complication for its benefit.

P2pool doesn't have a genesis block exactly, so there is a risk that the chain would be replaced and miners would lose their stored money.  The trick is to keep that incentive low.

[kwukduck]

I've been mining at p2pool for months now, without problems, but over the past week i've seen my payout plummet, today it has arrived at 0.00000.
The console tells me everything is fine, it's connected to both bitcoin and the peers, it's running at X GH/s steady, shares are being accepted and i have it running 24/7, yet it tells me 'current payout: 0.00000'.

Whats going on here?

[kjj]

I've been mining at p2pool for months now, without problems, but over the past week i've seen my payout plummet, today it has arrived at 0.00000.
The console tells me everything is fine, it's connected to both bitcoin and the peers, it's running at X GH/s steady, shares are being accepted and i have it running 24/7, yet it tells me 'current payout: 0.00000'.

Whats going on here?

Your hashrate is not capable of keeping a share active at all times.  See "Expected time to share:"

[kwukduck]

I've been mining at p2pool for months now, without problems, but over the past week i've seen my payout plummet, today it has arrived at 0.00000.
The console tells me everything is fine, it's connected to both bitcoin and the peers, it's running at X GH/s steady, shares are being accepted and i have it running 24/7, yet it tells me 'current payout: 0.00000'.

Whats going on here?

Your hashrate is not capable of keeping a share active at all times.  See "Expected time to share:"

Hmm seems you're right, expected time to share 9.5 hours.
Any idea what could cause this, or better yet, how to resolve the issue?
I'm using a bfl single. cgminer tells me shares get accepted continuously but the p2pool console says 0 shares.

[kjj]

Resolve it?  Just wait.  Maybe check the graphs to make sure your miner graphs are mostly blue and not red.

A share reported by your miner is not the same as a p2pool share.

[IYFTech]

OK, can't afford to keep the node running or mining at p2pool any longer with the current lack of shares issue for smaller (70Gig?) miners - so the node is off for good I'm afraid. It was fun while it lasted though. Maybe if there's a change to the program I'll switch on again, we'll see.......

Adios 

[gmaxwell]

I'm not sure what your issue is there. Low hash-rate miners won't be paid in every single block— once their rate is low enough that they don't constantly have a share in the window, but when they do get paid you'll be "overpaid" and as a result there returns will be as expected on average.

[Schleicher]

Low hash-rate miners won't be paid in every single block

Lol, yeah, that's true:

Local rate: 322MH/s (0.0% DOA) Expected time to share: 63.5 days

If we have bad luck it will be half a year.

[IYFTech]

I'm not sure what your issue is there. Low hash-rate miners won't be paid in every single block— once their rate is low enough that they don't constantly have a share in the window, but when they do get paid you'll be "overpaid" and as a result there returns will be as expected on average.

I wouldn't call 70+gig a low rate miner - not yet anyway. I just got tired of watching blocks fly by without finding shares, therefore no payments. It's false economy.

But I don't have an issue with it, I'll use my hash rate on another pool, one that pays for the work 

[Krak]

Low hash-rate miners won't be paid in every single block

Lol, yeah, that's true:

Local rate: 322MH/s (0.0% DOA) Expected time to share: 63.5 days

If we have bad luck it will be half a year.

I wouldn't call that a "low hashrate" miner. That's more of a micro hashrate.

[Carlton Banks]

I'm not sure what your issue is there. Low hash-rate miners won't be paid in every single block— once their rate is low enough that they don't constantly have a share in the window, but when they do get paid you'll be "overpaid" and as a result there returns will be as expected on average.

I wouldn't call 70+gig a low rate miner - not yet anyway. I just got tired of watching blocks fly by without finding shares, therefore no payments. It's false economy.

But I don't have an issue with it, I'll use my hash rate on another pool, one that pays for the work 

I think what you mean is: "I'll use my hash rate on another pool, one that pays less for the work at a commensurately lower share threshold, and I'll end up with the same amount of reward long term  "

[ksenter]

It is,  however, really disheartening to get 0 accepted shares for days.  My expected time to share is 15.4 hours but it's been days since I got one.  I did get an orphaned share a couple of days ago.  I can't really complain too much because the prior week I was getting shares way more frequent than the expected time to share.  But it still sucks watching it sit at 0 for so long...

[xyzzy099]

I'm not sure what your issue is there. Low hash-rate miners won't be paid in every single block— once their rate is low enough that they don't constantly have a share in the window, but when they do get paid you'll be "overpaid" and as a result there returns will be as expected on average.

I wouldn't call 70+gig a low rate miner - not yet anyway. I just got tired of watching blocks fly by without finding shares, therefore no payments. It's false economy.

But I don't have an issue with it, I'll use my hash rate on another pool, one that pays for the work 

I think what you mean is: "I'll use my hash rate on another pool, one that pays less for the work at a commensurately lower share threshold, and I'll end up with the same amount of reward long term  "

I think what he is really saying is "I want to mine on a pool that has less variance for me."  Seems like a reasonable move to me.  Solo mining also pays the same in the long run, but most people mine on pools to reduce variance.

Also, some other pools also have no fees, and pay transaction fees to miners, just like p2pool.

Don't get me wrong - I mine on p2pool, and I wish everyone did, just so over-centralization of mining would no longer be a threat.  But I definitely understand the frustration of small miners with the variance they see on p2pool.

[xyzzy099]

I mine on p2pool, and I wish everyone did, just so over-centralization of mining would no longer be a threat.

If over-centralization of mining would no longer be a threat if everyone mined on p2pool, why don't we just force everyone to mine on p2pool? Why don't we just change the protocol to the p2pool protocol?

(Hint: I'm not actually suggesting we do this.)

Then why bother asking the question?  Maybe I need a bigger hint to see what your point is.  And who is this 'we' you speak of?