p2pool - Decentralized, Absolutely DoS-Proof, Pool Hopping-Proof Pool [archival]

[EskimoBob]

Testing now. This is really exciting for some odd reason
Let me ask this: If this is going to work, can we say "Cheerio, Pool Lords!" and forget all about the mining pools, pool-hopping etc?

[1714187228]

1714187228

[1714187228]

1714187228

[1714187228]

1714187228

[1714187228]

1714187228

[1714187228]

1714187228

[1714187228]

1714187228

[Meni Rosenfeld]

Let me ask this: If this is going to work, can we say "Cheerio, Pool Lords!" and forget all about the mining pools, pool-hopping etc?

No, p2pool will probably be more suitable for large miners and normal pools than for at-home miners. But it will definitely help (eg, allow mining for a small pool without much variance).

[BurningToad]

Testing now. This is really exciting for some odd reason
Let me ask this: If this is going to work, can we say "Cheerio, Pool Lords!" and forget all about the mining pools, pool-hopping etc?

I hope so... (and I run a pool!)

It also seems that I am the majority of the P2P pool at the moment, cmon, need more MH/s!

[nanotube]

Let me ask this: If this is going to work, can we say "Cheerio, Pool Lords!" and forget all about the mining pools, pool-hopping etc?

No, p2pool will probably be more suitable for large miners and normal pools than for at-home miners. But it will definitely help (eg, allow mining for a small pool without much variance).

there can be many p2pools of larger or smaller size... so if "the main p2pool" gets too big for small miners to see reliable payouts, smaller ones can bud off. so yes, i indeed can foresee that the p2pool framework will overtake centralized pools.

[jgarzik]

I'm working on this on the wiki - http://wiki.bitcoin-otc.com/wiki/P2Pool.

Why is it on the bitcoin-otc wiki instead of the Bitcoin wiki?

Agreed, the main bitcoin wiki is what I had expected...

[grue]

Myself, um, dude.

No pony sigs. kthx.

not sure if serious or joking

[Meni Rosenfeld]

Let me ask this: If this is going to work, can we say "Cheerio, Pool Lords!" and forget all about the mining pools, pool-hopping etc?

No, p2pool will probably be more suitable for large miners and normal pools than for at-home miners. But it will definitely help (eg, allow mining for a small pool without much variance).

there can be many p2pools of larger or smaller size... so if "the main p2pool" gets too big for small miners to see reliable payouts, smaller ones can bud off. so yes, i indeed can foresee that the p2pool framework will overtake centralized pools.

At the very least you'll need someone to run a full network node for you, since eventually it will probably be impossible at home. So he may as well act as a pool with monitoring features etc.

[twmz]

I tried this last night and didn't have much success.  Here is the sequence I saw a couple times before I gave up:

* Start local bitcoind
* Startup p2pool python code with the user and password to the local bitcoind and with the -a parameter to specify an address from my real wallet (the local bitcoind does not have my real wallet and I don't want it to hold funds since it is not secured)
* Start poclbm miner (up to date from git)
* poclbm starts mining ok initially
* At some point, in the p2pool kind of goes insane and there is a long flurry of back to back messages about "generating".
* At about the same point or slightly after, the miner starts reporting RPC problems.  'getwork' requests are timing out and not returning.  Restarting the miner does not help.
* The "generating" flurry stops and the output of p2pool goes back to normally looking debug spam, but the miner never seems to be able to 'getwork' again and is stuck.
* Restarting p2pool makes things start working again for another minute or two

I apologize that I don't have the actual output because I didn't save it.  If this is not a known problem and isn't happening to others, let me know, and I can try it again later and capture the verbose output of both the miner and of p2pool.

[Shevek]

(...)
* poclbm starts mining ok initially
* At some point, in the p2pool kind of goes insane and there is a long flurry of back to back messages about "generating".

poclbm had a bug past night dealing with new difficulty. Try again with fresh poclbm and fresh p2pool. BTW, forrestv said that launching the program with python -O saves a lot of output.

[burp]

Um, p2pool has 100% cpu usage for me. Anyone else?

And right now I get spammed with

Code:

end generate_transaction
start generate_transaction
end generate_transaction
start generate_transaction
end generate_transaction
start generate_transaction

messages...

[EskimoBob]

Um, p2pool has 100% cpu usage for me. Anyone else?

And right now I get spammed with

Code:

end generate_transaction
start generate_transaction
end generate_transaction
start generate_transaction
end generate_transaction
start generate_transaction

messages...

Use: python -O run_p2pool.py username password

[burp]

Um, p2pool has 100% cpu usage for me. Anyone else?

And right now I get spammed with

Code:

end generate_transaction
start generate_transaction
end generate_transaction
start generate_transaction
end generate_transaction
start generate_transaction

messages...

Use: python -O run_p2pool.py username password

This doesn't help with the cpu usage problem. Also p2pool eats more and more memory. I stopped it at 600MB now. It didn't  even react to SIGINT anymore and I had to SIGKILL it.

[twmz]

Update:  

I just tried it again now and had the same problem.  In particular, I noticed errors about requests to bitcoind (http://127.0.0.1:8332) timing out (getwork requests and getting tx details, if I recall).

I was running this version of bitcoind: https://github.com/forrestv/bitcoin/tree/getrawtransaction because I thought it would be beneficial to have the rawtransaction support (side note: isn't our mining kind of useless to the network if we never include other people's transactions in our found blocks?  maybe I am misunderstanding the functionality added when getrawtransaction is present).

In any case, I tried switching to the officiall 0.3.24 release version of bitcoind and all of the problems went away.  The timeout messages are gone and my miner is now reliably mining and finding shares.

[jgarzik]

Where does p2pool get its transaction details, in order to rebuild the merkle tree with a new coinbase + block header?

That is not clear to me, from either this thread or the wiki.

A better technical description would be greatly appreciated.

[Shevek]

Following instructions at http://wiki.bitcoin-otc.com/wiki/P2Pool (Downloaded from Git repository), I start bitcoind (latest Linux version) and when starting p2pool, I get an error -10. What does that mean?

netstat -a shows that bitcoind is listening on ports 8332 and 8333.

Sometimes I've got strange errors when I launched p2pool quick after bitcoind. If this is your case, simply wait a couple of minutes.

[Shevek]

Sometimes I've got strange errors when I launched p2pool quick after bitcoind. If this is your case, simply wait a couple of minutes.

Okay, that doesn't seem to make a difference here. P2Pool fails consistently in main.py, line 31 (getwork). I can telnet to localhost:8332 just fine. I've also setup port forwarding and firewall rules for ports 8333 and 9333 in router and local machine.

Sorry for the trivial advice but... did you set server=1 in your bitcoin.conf?

[twmz]

Is there a known problem with poclbm (latest git) not checking the actual target of a getwork?  I don't see error messages about submitted shares being rejected for not being below the target, but I am also finding shares about the same rate as I do normal pool serving up exclusivly difficulty 1 shares.  I can clearly see in my proxy logs that the difficulty coming back in getwork requests is not 1. And I can see that the shares being found are are getworks with higher than difficulty 1 targets. The average difficulty of of the getworks associated with found shares over the past couple hours has been 3.  But I have found the same number of shares as I do when mining exclusively at difficulty 1.  Something doesn't seem right.

[twmz]

Is there a known problem with poclbm (latest git) not checking the actual target of a getwork?

Ok, I have confirmed that poclbm will happily submit shares that don't meat the target requirements.  This explains the shares/hour being identical to past experience even though share difficulty is 3-4x higher with p2pool.

My next question is does p2pool correctly reject these shares?  I can tell that p2pool never tells the miner they are rejected (the "result" of the share submission is always "1" (true)), but there may be some place else in the p2pool code that notices the hash does not meet the target requirements.  I'm not that good at reading python code, so I can't tell by looking.

[forrestv]

* The "generating" flurry stops and the output of p2pool goes back to normally looking debug spam, but the miner never seems to be able to 'getwork' again and is stuck.
* Restarting p2pool makes things start working again for another minute or two

Um, p2pool has 100% cpu usage for me. Anyone else?

If it's using a large amount of CPU, it's validating the sharechain, which currently can take a few minutes and a lot of memory. I'm going to change this to be more verbose and more efficient.

Where does p2pool get its transaction details, in order to rebuild the merkle tree with a new coinbase + block header?

It listens for transactions being broadcast from the local bitcoind via its P2P interface.

[Un zafado cualquiera]

Test in progress. good luck everyone.