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

[theonewhowaskazu]

The biggest problem with p2pool is the fact that it reduces my Jupiter's hashrate SIGNIFICANTLY. Hopefully either KNC or P2pool devs will fix that soon.

[1713494414]

1713494414

[1713494414]

1713494414

[1713494414]

1713494414

[theonewhowaskazu]

Hello again p2poolers 

OK, decided to switch my node on again after upgrading to 13.10. I'm struggling with getting slush's stratum proxy to play with p2pool so thought I'd ask here - does it work with p2pool? If so, can someone give me a heads up?

Ta muchly 

Anyone? 

From my experience, it doesn't seem to work with p2pool at all. The BFG one will work, though.

[IYFTech]

Hello again p2poolers 

OK, decided to switch my node on again after upgrading to 13.10. I'm struggling with getting slush's stratum proxy to play with p2pool so thought I'd ask here - does it work with p2pool? If so, can someone give me a heads up?

Ta muchly 

Anyone? 

From my experience, it doesn't seem to work with p2pool at all. The BFG one will work, though.

Thanks for the answer theonewhowaskazu.

I used to use the bfgminer one, but it does not support longpoll & I was getting over 30% DOA/stale/rejects with p2pool. Is there a reason why it works with every other pool perfectly well but not p2pool? Seems a bit strange eh?

Peace.

[alanwgeorge]

I get these quite often.  Is this from the miner connections or the Bitcoin-Qt connection?  Is it something I should worry about?

2014-01-16 18:59:38.826198 > Unhandled error in Deferred:
2014-01-16 18:59:38.826283 > Unhandled Error
2014-01-16 18:59:38.826316 > Traceback (most recent call last):
2014-01-16 18:59:38.826344 > Failure: twisted.internet.defer.TimeoutError: in ReplyMatcher

Thanks!

[int03h]

I get these quite often.  Is this from the miner connections or the Bitcoin-Qt connection?  Is it something I should worry about?

2014-01-16 18:59:38.826198 > Unhandled error in Deferred:
2014-01-16 18:59:38.826283 > Unhandled Error
2014-01-16 18:59:38.826316 > Traceback (most recent call last):
2014-01-16 18:59:38.826344 > Failure: twisted.internet.defer.TimeoutError: in ReplyMatcher

Thanks!

https://github.com/blixnood/P2PoolExtendedFrontEnd fixes this.

I dont know why .. it just does.

[IYFTech]

Hello again p2poolers  

OK, decided to switch my node on again after upgrading to 13.10. I'm struggling with getting slush's stratum proxy to play with p2pool so thought I'd ask here - does it work with p2pool? If so, can someone give me a heads up?

Ta muchly  

Anyone?  

From my experience, it doesn't seem to work with p2pool at all. The BFG one will work, though.

Thanks for the answer theonewhowaskazu.

I used to use the bfgminer one, but it does not support longpoll & I was getting over 30% DOA/stale/rejects with p2pool. Is there a reason why it works with every other pool perfectly well but not p2pool? Seems a bit strange eh?

Peace.

Obviously not then.

Oh well, off goes the node again - just spent a day upgrading everything to 13.10 for my new hardware in an effort to run my node again and guess what - it's not compatible with p2pool. Again.

Back to the Guild it is. Adios p2poolers  

[HellDiverUK]

You should probably stop feeding the troll.  He'll never use p2pool long enough or with any objectivity.  He just knows he doesn't like it without having a valid reason.

I ran p2pool for 4 months when I was mining with GPUs.  It worked fine then.  Heck, I had a public node with other miners using it 24/7. 

But, that was 6 months ago. I stopped using p2pool because it couldn't cope with my shiny new ASICMiner Blades.  I left hoping things would be fixed and I'd be back.

Time has moved on, p2pool has been left back in August 2013 with all sorts of issues, which the devs won't or can't fix. They do the "it's someone else's fault" thing that every other software developer defaults to when they don't have a clue.

I still love the idea of p2pool, but its SO FUCKING FRUSTRATING that it isn't being fixed so it WORKS.  How can it be expected to become more commonly used when it's so frigging wobbly and fussy? 

It's no wonder so many people flock to BTCGuild and Ghash.io - they both just work, no messing, no fuss.

[K1773R]

I was curious why all the p2pool payouts are listed on blockchain's "Strange Transactions" list.

https://blockchain.info/strange-transactions

"Unable to decode input address" 0 BTC

this txout is for identifying blocks of p2pool for internal funcitonality (ie, found block in last 24h and others)

[nreal]

You should probably stop feeding the troll.  He'll never use p2pool long enough or with any objectivity.  He just knows he doesn't like it without having a valid reason.

I ran p2pool for 4 months when I was mining with GPUs.  It worked fine then.  Heck, I had a public node with other miners using it 24/7. 

But, that was 6 months ago. I stopped using p2pool because it couldn't cope with my shiny new ASICMiner Blades.  I left hoping things would be fixed and I'd be back.

Time has moved on, p2pool has been left back in August 2013 with all sorts of issues, which the devs won't or can't fix. They do the "it's someone else's fault" thing that every other software developer defaults to when they don't have a clue.

I still love the idea of p2pool, but its SO FUCKING FRUSTRATING that it isn't being fixed so it WORKS.  How can it be expected to become more commonly used when it's so frigging wobbly and fussy? 

It's no wonder so many people flock to BTCGuild and Ghash.io - they both just work, no messing, no fuss.

I stink that p2pool works like a charm, the only bad thing in it is share difficulty, if it could be divided to 1/10 from what it is now.... There would be alot of happy people smiling here and there ?

 

[alanwgeorge]

I get these quite often.  Is this from the miner connections or the Bitcoin-Qt connection?  Is it something I should worry about?

2014-01-16 18:59:38.826198 > Unhandled error in Deferred:
2014-01-16 18:59:38.826283 > Unhandled Error
2014-01-16 18:59:38.826316 > Traceback (most recent call last):
2014-01-16 18:59:38.826344 > Failure: twisted.internet.defer.TimeoutError: in ReplyMatcher

Thanks!

https://github.com/blixnood/P2PoolExtendedFrontEnd fixes this.

I dont know why .. it just does.

Whatever "magic" happened for you it didn't happen for me.  I've had P2PoolExtendedFrontEnd installed for a while and still seeing these timeout errors.

2014-01-17 08:49:06.336406 > Unhandled error in Deferred:
2014-01-17 08:49:06.336631 > Unhandled Error
2014-01-17 08:49:06.336702 > Traceback (most recent call last):
2014-01-17 08:49:06.336760 > Failure: twisted.internet.defer.TimeoutError: in ReplyMatcher

What does "ReplyMatcher" do?

[forrestv]

Whatever "magic" happened for you it didn't happen for me.  I've had P2PoolExtendedFrontEnd installed for a while and still seeing these timeout errors.

2014-01-17 08:49:06.336406 > Unhandled error in Deferred:
2014-01-17 08:49:06.336631 > Unhandled Error
2014-01-17 08:49:06.336702 > Traceback (most recent call last):
2014-01-17 08:49:06.336760 > Failure: twisted.internet.defer.TimeoutError: in ReplyMatcher

What does "ReplyMatcher" do?

ReplyMatcher matches replies to messages sent to bitcoind with responses. It's only used for the get_block_header message, and this error means that bitcoind didn't reply ... either it timed out or bitcoind doesn't have the header in question. So it could suggest that your Bitcoin node is slow or not up-to-date.. but seeing these rarely is normal.

[alanwgeorge]

Whatever "magic" happened for you it didn't happen for me.  I've had P2PoolExtendedFrontEnd installed for a while and still seeing these timeout errors.

2014-01-17 08:49:06.336406 > Unhandled error in Deferred:
2014-01-17 08:49:06.336631 > Unhandled Error
2014-01-17 08:49:06.336702 > Traceback (most recent call last):
2014-01-17 08:49:06.336760 > Failure: twisted.internet.defer.TimeoutError: in ReplyMatcher

What does "ReplyMatcher" do?

ReplyMatcher matches replies to messages sent to bitcoind with responses. It's only used for the get_block_header message, and this error means that bitcoind didn't reply ... either it timed out or bitcoind doesn't have the header in question. So it could suggest that your Bitcoin node is slow or not up-to-date.. but seeing these rarely is normal.

I'm seeing 2 or 3 of these timeouts a minute consistently.  I assume that doesn't qualify as rarely.

I'm running Bitcoin-Qt (v0.8.6-beta I believe thats the latest, sum on mine and the latest download are same) and p2pool on a Mac Mini.  8333, 9333 and 9332 ports are forwarded.  Miners are on a RaspberryPi running bfgminer.  Network is WiFi.  My GetBlockTemplate latency mean for the last hour is .197s.

Anything specific I should look at?

Thanks!

[PatMan]

Your wifi......ditch it. 

[alanwgeorge]

Whatever "magic" happened for you it didn't happen for me.  I've had P2PoolExtendedFrontEnd installed for a while and still seeing these timeout errors.

2014-01-17 08:49:06.336406 > Unhandled error in Deferred:
2014-01-17 08:49:06.336631 > Unhandled Error
2014-01-17 08:49:06.336702 > Traceback (most recent call last):
2014-01-17 08:49:06.336760 > Failure: twisted.internet.defer.TimeoutError: in ReplyMatcher

What does "ReplyMatcher" do?

ReplyMatcher matches replies to messages sent to bitcoind with responses. It's only used for the get_block_header message, and this error means that bitcoind didn't reply ... either it timed out or bitcoind doesn't have the header in question. So it could suggest that your Bitcoin node is slow or not up-to-date.. but seeing these rarely is normal.

I'm seeing 2 or 3 of these timeouts a minute consistently.  I assume that doesn't qualify as rarely.

I'm running Bitcoin-Qt (v0.8.6-beta I believe thats the latest, sum on mine and the latest download are same) and p2pool on a Mac Mini.  8333, 9333 and 9332 ports are forwarded.  Miners are on a RaspberryPi running bfgminer.  Network is WiFi.  My GetBlockTemplate latency mean for the last hour is .197s.

Anything specific I should look at?

Thanks!

Your wifi......ditch it.  

Which link?  

Miners to p2pool
bitcoind to world

I have run the miners locally on the Mac Mini and both bitcoind and p2pool, that did seem to make a difference on the timeouts.

From forrestv response, the ReplyMatcher timeouts are between bitcoind and p2pool.  So WiFi should not effect, I guess.

My latency and efficiency seem within norms so may be the ReplyMatcher timeouts are just noise.

[smooth]

Good point about the average.  

However, when you factor in the additional variance introduced by high share difficulty and reject rate, and you get more variance than you would elsewhere.  

Yes?

You get more variance. Full stop. P2pool has higher variance than other pools. The compensation is higher expected earnings from no pool fees, payout of transaction fees, and no risk of being cheated by the pool.

The question most of us are asking is whether it makes sense to make it an absolute priority to minimize variance. There is no objective answer, but in real terms the variance for a small miner who expects to pull $2/day at current difficulty is not "large" or "extreme." It is large relative to earnings, but the earnings are tiny, and so is the variance. Even then, it has been explained that you will be getting payouts at that tiny hash rate about 50% of the time. Not too bad.

The main thing you get by focusing intensely on minimizing variance is a nice pretty graph of flat payouts. If that is really important enough to you that you are willing to forgo higher expected earnings, go for it. No one can say objectively that's the wrong thing to do, but we can question it.

BTW, the reject rate doesn't really increase variance though, it just (slightly) reduces the share rate. A share that is rejected is exactly the same as a share that you never found. Other than that the high rejects might be telling you there is something wrong with your setup.

[PatMan]

Whatever "magic" happened for you it didn't happen for me.  I've had P2PoolExtendedFrontEnd installed for a while and still seeing these timeout errors.

2014-01-17 08:49:06.336406 > Unhandled error in Deferred:
2014-01-17 08:49:06.336631 > Unhandled Error
2014-01-17 08:49:06.336702 > Traceback (most recent call last):
2014-01-17 08:49:06.336760 > Failure: twisted.internet.defer.TimeoutError: in ReplyMatcher

What does "ReplyMatcher" do?

ReplyMatcher matches replies to messages sent to bitcoind with responses. It's only used for the get_block_header message, and this error means that bitcoind didn't reply ... either it timed out or bitcoind doesn't have the header in question. So it could suggest that your Bitcoin node is slow or not up-to-date.. but seeing these rarely is normal.

I'm seeing 2 or 3 of these timeouts a minute consistently.  I assume that doesn't qualify as rarely.

I'm running Bitcoin-Qt (v0.8.6-beta I believe thats the latest, sum on mine and the latest download are same) and p2pool on a Mac Mini.  8333, 9333 and 9332 ports are forwarded.  Miners are on a RaspberryPi running bfgminer.  Network is WiFi.  My GetBlockTemplate latency mean for the last hour is .197s.

Anything specific I should look at?

Thanks!

Your wifi......ditch it. 

Which link? 

Miners to p2pool
bitcoind to world

I have run the miners locally on the Mac Mini and both bitcoind and p2pool, that did seem to make a difference on the timeouts.

From forrestv response, the ReplyMatcher timeouts are between bitcoind and p2pool.  So WiFi should not effect, I guess.

My latency and efficiency seem within norms so may be the ReplyMatcher timeouts are just noise.

I would get rid of all of it & go wired. Mining over wifi is notorious for causing errors/problems, it's the weakest link & by removing it you are eliminating it as the cause of your problem.

[mdude77]

You get more variance. Full stop. P2pool has higher variance than other pools. The compensation is higher expected earnings from no pool fees, payout of transaction fees, and no risk of being cheated by the pool.

What you say is true, except it really doesn't equate to "higher expected earnings".  Yes you can use p2pool with 0 fees, but last I checked, it's frowned upon.  Most, if not all pools, now pay transaction fees.  And risk of being cheated by a pool is there, but slim.  Most pools are well established, we know who is trust worthy and who isn't.

The main thing you get by focusing intensely on minimizing variance is a nice pretty graph of flat payouts. If that is really important enough to you that you are willing to forgo higher expected earnings, go for it. No one can say objectively that's the wrong thing to do, but we can question it.

Again, there is no higher expected earnings, unless you purposely choose not to "donate" to forrest.

BTW, the reject rate doesn't really increase variance though, it just (slightly) reduces the share rate. A share that is rejected is exactly the same as a share that you never found. Other than that the high rejects might be telling you there is something wrong with your setup.

That's like saying if you aren't mining, it's the same as getting a reject.  Not sure what your point is?  The pool right now is at 19% rejects.  You can say all you want about "something wrong with your setup", just realize it's everyone's setup, and certainly not mine, since I'm not in p2pool.  

I stand by my statement, high reject rate increases variance.  If that precious share rejects after my 4 day average time to find a share, then, on average, it's going to be another 4 days before I get another chance.  And then, there's a 19% chance it'll reject too!  "normal" pools average around 2% reject rate.   That means my chances of getting a reject are 9.5x higher on p2pool than on a conventional pool.  How does that not increase variance?

M

[Krak]

Which link? 

Miners to p2pool
bitcoind to world

I have run the miners locally on the Mac Mini and both bitcoind and p2pool, that did seem to make a difference on the timeouts.

From forrestv response, the ReplyMatcher timeouts are between bitcoind and p2pool.  So WiFi should not effect, I guess.

My latency and efficiency seem within norms so may be the ReplyMatcher timeouts are just noise.

I would get rid of all of it & go wired. Mining over wifi is notorious for causing errors/problems, it's the weakest link & by removing it you are eliminating it as the cause of your problem.

Though if you have to choose one, I would definitely get the p2pool host on ethernet. Both would be best, but it's slightly less important for the miners.

[smooth]

You get more variance. Full stop. P2pool has higher variance than other pools. The compensation is higher expected earnings from no pool fees, payout of transaction fees, and no risk of being cheated by the pool.

What you say is true, except it really doesn't equate to "higher expected earnings".  Yes you can use p2pool with 0 fees, but last I checked, it's frowned upon.

Frowned upon by whom? There is no pool operator to whom you would be paying fees.

Most, if not all pools, now pay transaction fees.

Very few pools pay transaction fees, merged mine for the benefit of the miner, and don't charge a fee. It costs money to run a pool, where do you think that money comes from?

And risk of being cheated by a pool is there, but slim.  Most pools are well established, we know who is trust worthy and who isn't.

You don't know with certainty who is untrustworthy until after they have been discovered to be untrustworthy (or incompetent). As you say the risk of being cheated is there, so the "expected" earnings are lower, even if the adverse event hasn't happened yet.

Again, there is no higher expected earnings, unless you purposely choose not to "donate" to forrest.

The higher expected earnings come from the combination of no fees, paying tx fees, merged mining, reduced orphan risk with p2pool, and no risk of cheating by the pool operator.

I stand by my statement, high reject rate increases variance.  If that precious share rejects after my 4 day average time to find a share, then, on average, it's going to be another 4 days before I get another chance.  And then, there's a 19% chance it'll reject too!

It doesn't really work that way. 80% of the time your share will be accepted! All this means is that instead of a 4 day average time (if that is the number), you have a 5 day average share time.

"normal" pools average around 2% reject rate.   That means my chances of getting a reject are 9.5x higher on p2pool than on a conventional pool.  How does that not increase variance?

These rejects don't mean the same thing. If the pool average is 19% (I didn't realize it was that high) and your reject rate is 19% then you haven't lost any earnings at all.  You are on par with all the other pool members so when it comes time to divide up the blocks you will get your proportional share.

[TechByPC]

Question:

With the PPLNS reward, how big is "N"? In other words, how long does someone need to be mining at a constant rate in order to receive their maximum reward for that hash rate?

Thanks.