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

[CrazyGuy]

I tried to drop my 2.5 gh/s minecart in, with the same version of phoenix but running the last release of linuxcoin.  I'm getting about 30% - 50% reject rate on all 7 gpus (3x5970s & 1x5830) Same miners pointing at Slush and I'm down at 1% reject rate. Does the P2P client have issues with nodes that have too many GPUs? Maybe I need to update the ATI drivers...

I run a single rig with 5 GPUs and phoenix and get about 5% reject.  Check your AGRESSION option.  It needs to be lower for p2pool in order to get lower stales.  I was getting 20-30% as well and dropped the AGRESSION from 9 down to 7 and reject rate dropped like a rock.  I don't know why this is necessary, but it seems to be.

Thanks, dropped aggression from 11 to 8 and no rejects yet. It did lower my hash rate by about 2.5% but that's better than 50% rejects.

[btc_artist]

Does bitcoind have to be caught up downloading the block chain before starting P2Pool?

[twmz]

Does bitcoind have to be caught up downloading the block chain before starting P2Pool?

Yes

[jsgarvin]

Does bitcoind have to be caught up downloading the block chain before starting P2Pool?

Yes

Is there a convenient way to ensure this happens unattended?

I'm not currently on p2pool.  I currently have my (linux) mining rig configured to run a script on boot that automatically gets cgminer going pointed at the right pools.  So, if the power goes out while I'm away, and then comes back on, everything (in theory), just fires right back up again.

So, if I were to switch to p2pool, I'd need some way of automatically launching bitcoind (I guess), monitoring it to see when it's caught up, and then launching the p2pool client.

Or, better yet, to make things easier on everyone, it would be nice if the p2pool client managed that itself so that one could launch it right away and it would monitor bitcoind and not start actually mining until the time was right.  But, not knowing what goes on between bitciond and p2pool, I admit that that might not be technically feasible.

[Red Emerald]

Does bitcoind have to be caught up downloading the block chain before starting P2Pool?

Yes

Is there a convenient way to ensure this happens unattended?

I'm not currently on p2pool.  I currently have my (linux) mining rig configured to run a script on boot that automatically gets cgminer going pointed at the right pools.  So, if the power goes out while I'm away, and then comes back on, everything (in theory), just fires right back up again.

So, if I were to switch to p2pool, I'd need some way of automatically launching bitcoind (I guess), monitoring it to see when it's caught up, and then launching the p2pool client.

Or, better yet, to make things easier on everyone, it would be nice if the p2pool client managed that itself so that one could launch it right away and it would monitor bitcoind and not start actually mining until the time was right.  But, not knowing what goes on between bitciond and p2pool, I admit that that might not be technically feasible.

I think its fine to launch p2pool even while bitcoind is out of date. Your shares will be rejected until it gets up to date, but once its up to date it should start working fine.

[Vanderbleek]

What would be really nice would be to bundle p2pool with bitcoin, so as long as the client (or bitcoind) is running p2pool would run. Not to mention providing it with a gui interface in bitcoinQT would probably help adoption.

[jsgarvin]

Am I correct in assuming that, when using the `-a` flag to specify a payout address, that the provided address need not be provided by the instance of bitcoind that p2pool is connected to? (i.e. it can be from another wallet on a completely different computer?)

[Red Emerald]

I just moved my main miner over to p2pool and it's running great.

Just as a backup, I am trying to install p2pool on another system.  It seems to be running, but I am getting a lot of failures. I get this over and over.

Code:

2012-01-30 23:53:24.088756 > Traceback (most recent call last):
2012-01-30 23:53:24.088990 > Failure: twisted.internet.error.ConnectionRefusedError: Connection was refused by other side: 111: Connection refused.
2012-01-30 23:53:25.091631 > Traceback (most recent call last):
2012-01-30 23:53:25.091863 > Failure: twisted.internet.error.ConnectionRefusedError: Connection was refused by other side: 111: Connection refused.
2012-01-30 23:53:26.094514 > Traceback (most recent call last):
2012-01-30 23:53:26.094820 > Failure: twisted.internet.error.ConnectionRefusedError: Connection was refused by other side: 111: Connection refused.
2012-01-30 23:53:27.097429 > Traceback (most recent call last):
2012-01-30 23:53:27.097661 > Failure: twisted.internet.error.ConnectionRefusedError: Connection was refused by other side: 111: Connection refused.
2012-01-30 23:53:29.547980 Processing 1001 shares from XXX.XXX.XXX.XXX:9334...
2012-01-30 23:53:29.859644 Requesting parent share c79e70a4 from XXX.XXX.XXX.XXX:9334
2012-01-30 23:53:29.860808 ... done processing 1001 shares. New: 1001 Have: 3006/~17280
2012-01-30 23:53:29.873999 > Traceback (most recent call last):

2012-01-30 23:54:13.473540 Processing 1001 shares from XXX.XXX.XXX.XXX:9334...
2012-01-30 23:54:40.850463 Requesting parent share 1c063c3c from XXX.XXX.XXX.XXX:9333
2012-01-30 23:54:40.851541 ... done processing 1001 shares. New: 1001 Have: 9016/~17280
2012-01-30 23:54:40.929094 Pool: 160GH/s in 9016 shares (375/9016 verified) Recent: 0.14% >218MH/s Shares: 0 (0 orphan, 0 dead) Peers: 4 (0 incoming)

Any ideas?  I'm running the current git head (5f49537)

EDIT: Ahaa! Namecoin wasn't running and I had it set to merge mine.  That error message could have been more helpful though.

[thirdlight]

Am I correct in assuming that, when using the `-a` flag to specify a payout address, that the provided address need not be provided by the instance of bitcoind that p2pool is connected to? (i.e. it can be from another wallet on a completely different computer?)

Yes.

Also, if anyone would like an address to test this with, I'll happily provide one 

[cheat_2_win]

I thought I had it figured out. But P2pool hasn't worked in my case:

Steps I followed are:

1. Ran Bitcoin client with RPC interface enabled (created bitcoin.conf and put 'rpcuser', 'rpcpassword' and 'server' in there). Let it update the block chain.
2. Ran run_p2pool.exe
3. Ran cgminer.exe with -o http://127.0.0.1:9332 -u <rpcuser> -p <rpcpassword>

run_p2pool seems to get stuck and 'cgminer.exe' complains: "No servers were found that could be used to get works from.".

Can anyone help?

[gnar1ta$]

I thought I had it figured out. But P2pool hasn't worked in my case:

Steps I followed are:

1. Ran Bitcoin client with RPC interface enabled (created bitcoin.conf and put 'rpcuser', 'rpcpassword' and 'server' in there). Let it update the block chain.
2. Ran run_p2pool.exe
3. Ran cgminer.exe with -o http://127.0.0.1:9332 -u <rpcuser> -p <rpcpassword>

run_p2pool seems to get stuck and 'cgminer.exe' complains: "No servers were found that could be used to get works from.".

Can anyone help?

It sounds like p2pool isn't connecting to bitcoind. You have to connect it by by using the username and password you set up in bitcoin.conf. Run_p2pool.exe <rpcuser> <rpcpass> and then connect cgminer using any user pass - no the ones you set up in bitcoin.conf.

[btc_artist]

I'm trying to get P2Pool going, but I'm running into problems:

1. bitcoind is running and downloading the blockchain. getinfo confirms it's currently at 140k blocks.
2. I run p2pool and get output like the following:

Code:

Install SciPy for more accurate confidence intervals!
2012-01-31 11:19:33.754597 p2pool (version e5e112f)
2012-01-31 11:19:33.754697
2012-01-31 11:19:33.841434 Testing bitcoind RPC connection to 'http://127.0.0.1:8332/' with username 'rpcuser'...
2012-01-31 11:19:38.846469 > Error while checking bitcoind identity:
2012-01-31 11:19:42.913352 > Traceback (most recent call last):
2012-01-31 11:19:42.913533 > Failure: twisted.internet.defer.TimeoutError: Getting http://127.0.0.1:8332/ took longer than 5 seconds.

I assume this means p2pool can't reach the bitcoind instance, but it's there and running.  Any ideas?

[m3ta]

I'm trying to get P2Pool going, but I'm running into problems:

1. bitcoind is running and downloading the blockchain. getinfo confirms it's currently at 140k blocks.
2. I run p2pool and get output like the following:

Code:

Install SciPy for more accurate confidence intervals!
2012-01-31 11:19:33.754597 p2pool (version e5e112f)
2012-01-31 11:19:33.754697
2012-01-31 11:19:33.841434 Testing bitcoind RPC connection to 'http://127.0.0.1:8332/' with username 'rpcuser'...
2012-01-31 11:19:38.846469 > Error while checking bitcoind identity:
2012-01-31 11:19:42.913352 > Traceback (most recent call last):
2012-01-31 11:19:42.913533 > Failure: twisted.internet.defer.TimeoutError: Getting http://127.0.0.1:8332/ took longer than 5 seconds.

I assume this means p2pool can't reach the bitcoind instance, but it's there and running.  Any ideas?

Just wait until the blockchain finishes downloading.

[btc_artist]

Well, I know I won't be able to mine until the block chain downloads completely, but I'm trying to figure out why p2pool can't talk to bitcoind.  I would assume it should be able to even if there are still blocks to download.

[cheat_2_win]

I thought I had it figured out. But P2pool hasn't worked in my case:

Steps I followed are:

1. Ran Bitcoin client with RPC interface enabled (created bitcoin.conf and put 'rpcuser', 'rpcpassword' and 'server' in there). Let it update the block chain.
2. Ran run_p2pool.exe
3. Ran cgminer.exe with -o http://127.0.0.1:9332 -u <rpcuser> -p <rpcpassword>

run_p2pool seems to get stuck and 'cgminer.exe' complains: "No servers were found that could be used to get works from.".

Can anyone help?

It sounds like p2pool isn't connecting to bitcoind. You have to connect it by by using the username and password you set up in bitcoin.conf. Run_p2pool.exe <rpcuser> <rpcpass> and then connect cgminer using any user pass - no the ones you set up in bitcoin.conf.

I actually used : run_p2pool.exe <rpcuser> <rpcpassword>

Initially, run_p2pool seems to be communicating with bitcoind. But after a while, there are no messages on the console.  On the cgminer console, first it says "Pool 0 not providing work fast enough", then after a while I get a message saying "Pool 0 http://127.0.0.1:9332 not responding". I checked the firewall. All the ports are opened appropriately. Bitcoin-qt has up-to-date block information as well.

[cheat_2_win]

Is there a recommended version of bitcoind to use with P2pool?

[Ente]

cheat:
please post the exact commands to run bitcoind and p2pool and their output.

Both of you:
You both seem to have problems connecting p2pool to the local bitcoind daemon.
(Seems like the windows I have available doesnt have telnet installed any more? lol)

Simply try http://127.0.0.1:8332 in your browser. If bitcoind is running, you will get a passwordprompt. You can check if your password is what you expect it to there. Entering the right credentials doesnt yield any information after that, though.

Trying the same with the p2pool daemon (127.0.0.1:9332) doesnt seem to give any output, btw.

Ente

[btc_artist]

Simply try http://127.0.0.1:8332 in your browser. If bitcoind is running, you will get a passwordprompt. You can check if your password is what you expect it to there. Entering the right credentials doesnt yield any information after that, though.

Yes. I get the password prompt, and entering the right credentials lets me past it.

[cheat_2_win]

cheat:
please post the exact commands to run bitcoind and p2pool and their output.

Both of you:
You both seem to have problems connecting p2pool to the local bitcoind daemon.
(Seems like the windows I have available doesnt have telnet installed any more? lol)

Simply try http://127.0.0.1:8332 in your browser. If bitcoind is running, you will get a passwordprompt. You can check if your password is what you expect it to there. Entering the right credentials doesnt yield any information after that, though.

Trying the same with the p2pool daemon (127.0.0.1:9332) doesnt seem to give any output, btw.

Ente

Output from http://127.0.0.1:8332 after I enter the username and password is : {"result":null,"error":{"code":-32700,"message":"Parse error"},"id":null}

At http://127.0.0.1:9332 sometimes give an output, other times it just takes forever to connect to.

I actually tried using 'python run_p2pool.py' as well as 'run_p2pool.exe' with appropriate rpc user name and password. Both have same problem.

So, what should be the output if I am able to connect to P2pool?

[forrestv]

Normal output is like this: http://im.forre.st/pb/07231883.txt

Can you pastebin the output that you're getting?