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

[zvs]

it may be slower, u wont get orphans due to that since u sibmit ur shares to the p2pool sharechain. i usually got 0 orphans and some doas due to the reason i got 100 connections to other nodes.

hmm, do you (or anyone else) have a list of all the p2pool nodes that are connectable (i guess there's an -addnode equivalent?) that sustain 3+ ghash??


oh, everyone feel free to addnode 5.9.24.81 in p2pool list

5.*.*.* is illegal.

if ur pool is accessable then just wait otherwise use --outgoing-conns
hashrate of the pool dosnt matter.

illegal?  it's just some hetzner server i've had for like 9 months =p

well, the higher the hashrate the more shares that originate from there was my thinking.  i'll try the outgoing thing later... thanks

[K1773R]

it may be slower, u wont get orphans due to that since u sibmit ur shares to the p2pool sharechain. i usually got 0 orphans and some doas due to the reason i got 100 connections to other nodes.

hmm, do you (or anyone else) have a list of all the p2pool nodes that are connectable (i guess there's an -addnode equivalent?) that sustain 3+ ghash??


oh, everyone feel free to addnode 5.9.24.81 in p2pool list

5.*.*.* is illegal.

if ur pool is accessable then just wait otherwise use --outgoing-conns
hashrate of the pool dosnt matter.

illegal?  it's just some hetzner server i've had for like 9 months =p

well, the higher the hashrate the more shares that originate from there was my thinking.  i'll try the outgoing thing later... thanks

this dosnt matter, shares are being broadcasted.

[zvs]

it may be slower, u wont get orphans due to that since u sibmit ur shares to the p2pool sharechain. i usually got 0 orphans and some doas due to the reason i got 100 connections to other nodes.

hmm, do you (or anyone else) have a list of all the p2pool nodes that are connectable (i guess there's an -addnode equivalent?) that sustain 3+ ghash??


oh, everyone feel free to addnode 5.9.24.81 in p2pool list

5.*.*.* is illegal.

if ur pool is accessable then just wait otherwise use --outgoing-conns
hashrate of the pool dosnt matter.

illegal?  it's just some hetzner server i've had for like 9 months =p

well, the higher the hashrate the more shares that originate from there was my thinking.  i'll try the outgoing thing later... thanks

this dosnt matter, shares are being broadcasted.

sure, but they aren't being broadcast fast enough across the network.   there wouldn't be so many orphans, if there wasnt a problem

i just had another share orphaned, that actually had a child (bitcoin address was 1Cgxxxxxxxxx)...  5 seconds later, both mine and 1Cg's were gone and replaced with another tree.

that's happened once approx. 2 1/2 months for the bitcoin blockchain  (the 2nd instance being on 10-07).  

maybe change the difficulty to even out to three per minute instead of the current... five, I think it is?

I'm at 72 shares, 9 orphans, 2 DOA..   my p2pool server is located on a major hub @ hetzner, I also get 10ms ping times to the OVH french datacenter....  bandwidth limits are 1gbit/1gbit.    i have a bitcoind that has 800 connections which was the 'hub' node on blockchain ever since it was implemented until a few days ago (it keeps dropping me and I don't run it 24/7 anymore).... so, i think it pushes things out fairly quickly.

the attitude people have about orphans on the p2pool network is troublesome.  it's not something that should exist to the extent that it does

[zvs]

I recompiled to do 30 outgoing connections and restarted, has been running about an hour or two so far...   nothing else has changed, didn't reload bitcoind or change anything else running on the server.  I guess I'll also mention that the pool DOA rate was ranging from 15-18% when I switched

Sample size isn't big enough now, but maybe in 5 or 6 hours

(you can see the swapover on my graph quite clearly @ the 9pm indicator)

[bitcoindaddy]

Is it necessary to use a bitcoin address on the command line that exists in the bitcoind instance you are running? For example, could you mine to a cloud based address (blockchain.info) that didn't exist on your local bitcoind?

[kjj]

Is it necessary to use a bitcoin address on the command line that exists in the bitcoind instance you are running? For example, could you mine to a cloud based address (blockchain.info) that didn't exist on your local bitcoind?

No.  Yes.

[zvs]

I recompiled to do 30 outgoing connections and restarted, has been running about an hour or two so far...   nothing else has changed, didn't reload bitcoind or change anything else running on the server.  I guess I'll also mention that the pool DOA rate was ranging from 15-18% when I switched

Sample size isn't big enough now, but maybe in 5 or 6 hours

(you can see the swapover on my graph quite clearly @ the 9pm indicator)

err, I killed it for now.  Have to redownload the blockchain, messed up while recompiling w/ higher fee settings.

I believe it was around 85 shares with 3 orphans.... with the 30 outgoing connections.   The precompiled version only allows you to set it to 10, however

ed: nm, found the stats file.  was 87 shares, 4 orphans and 3 DOA.... vs the 72 shares, 9 orphans, and 2 DOA earlier.  the DOA was probably just random chance.  don't think so for orphans

[bitcoindaddy]

Is it necessary to use a bitcoin address on the command line that exists in the bitcoind instance you are running? For example, could you mine to a cloud based address (blockchain.info) that didn't exist on your local bitcoind?

No.  Yes.

Thanks.

I'm trying to like p2pool, but I'm getting 50% rejects. I have a good cable internet connection with nothing else significant running on it.

Does it matter if I run bitcoin-qt vs. bitcoind?

Here's my cgminer config:

Code:

{
"pools" : [
        {
                "url" : "http://127.0.0.1:9332",
                "user" : "none",
                "pass" : "none"
        }

]
,
"intensity" : "7,7,7",
"vectors" : "2,2,2",
"worksize" : "128,128,128",
"kernel" : "diablo,diablo,diablo",
"gpu-engine" : "800-875,800-875,800-875",
"gpu-fan" : "0-85,0-85,0-85",
"gpu-memclock" : "300,300,300",
"gpu-memdiff" : "0,0,0",
"gpu-powertune" : "0,0,0",
"gpu-vddc" : "0.000,0.000,0.000",
"temp-cutoff" : "95,95,95",
"temp-overheat" : "85,85,85",
"temp-target" : "75,75,75",
"failover-only" : true,
"auto-fan" : true,
"auto-gpu" : true,
"expiry" : "120",
"gpu-dyninterval" : "7",
"gpu-platform" : "0",
"gpu-threads" : "2",
"log" : "5",
"queue" : "1",
"retry-pause" : "5",
"scan-time" : "60",
"temp-hysteresis" : "3",
"shares" : "0",
"kernel-path" : "/usr/local/bin"

[rav3n_pl]

Try change to:

Code:

"gpu-threads" : "1",
"queue" : "0",

What is your GBT latency? 50% is soo high stale.

[bitcoindaddy]

Try change to:

Code:

"gpu-threads" : "1",
"queue" : "0",

What is your GBT latency? 50% is soo high stale.

Changing queue to zero cut my rejects down to about 6 percent. I have not tried changing the threads because that will halve my speed, right?

Where do I find GBT latency?

[rdponticelli]

i have a bitcoind that has 800 connections

This may be the cause of your DOA. It was already said here, but you shouldn't have that many connections on your bitcoind. The network code for bitcoin isn't yet optimized and with that many connections you're effectively harming yourself and the network as a whole...

[bitcoindaddy]

Try change to:

Code:

"gpu-threads" : "1",
"queue" : "0",

What is your GBT latency? 50% is soo high stale.

Changing queue to zero cut my rejects down to about 6 percent. I have not tried changing the threads because that will halve my speed, right?

Where do I find GBT latency?

OK, in the graphs I found - "getwork latency mean 0.787s".

Another thing that has me worried is the "Expected after mining for 24 hours" is only half of what I get on a PPS server.

[PatMan]

I was just looking at the charts and was hit by several clumps of tumbleweed..........

[Miner99er]

I was just looking at the charts and was hit by several clumps of tumbleweed..........

And as if by magic... Cha-ching.

[forrestv]

Another thing that has me worried is the "Expected after mining for 24 hours" is only half of what I get on a PPS server.

It says:

Code:

Payout if a block were found NOW: x.yz BTC to 1AfPXDktciQyw1Q1wTwT4xPZTbhAKruyL7. Expected after mining for 24 hours: x.yz BTC

So it means the expected payout in the next P2Pool block, not the expected daily revenue! We expect to get a block every 12 hours, so this number should be about half your calculated daily revenue.

[K1773R]

"2d 21h 53m" what a long fking block...

[PatMan]

I was just looking at the charts and was hit by several clumps of tumbleweed..........

And as if by magic... Cha-ching.

I'll have to make more inane comments more often.....

[PatMan]

"2d 21h 53m" what a long fking block...

Longest one since I joined this pool. Cripes man.

[K1773R]

"2d 21h 53m" what a long fking block...

Longest one since I joined this pool. Cripes man.

and our luck stats went down like shit
it could have been worse (longer or stale)

[rav3n_pl]

Changing queue to zero cut my rejects down to about 6 percent. I have not tried changing the threads because that will halve my speed, right?

Wrong
It will change number of threads on GPU, it MIGHT reduce for about 5% but not always.