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

[coinme.info]

GBT latency is the big issue for me as I'm 12,000+km of fiber away from Matt's US-West node which I'm using as it was 100ms faster than the AU node.
This is normal as any US-West servers are normally lower latency than Asia nodes due to network topology.

Currently GBT is between 0.2 - 0.4 so I'm happy with that and the low DOA and normal shares being found seems to support that things are running ok.

This on 7200 RPM Sata as the SSD died.

Where are you? Would a local node help?

New Zealand

A Sydney or Auckland node would help the AUS/NZ region for sure.

You can see the routes below

First the public.au.relay.mattcorallo.com


      Host                                           -   %  | Sent | Recv | Best | Avrg | Wrst | Last |
                           fw                          -    0 |   51 |   51 |    0 |    0 |    2 |    1 |
               xxxxxxxx.jetstream.xtra.co.nz  -    3 |   48 |   47 |   14 |   24 |   57 |   17 |
mdr-ip20e-int.msc.global-gateway.net.nz -    0 |   51 |   51 |   21 |   32 |   62 |   22 |
      ae0-10.akbr6.global-gateway.net.nz -    3 |   48 |   47 |   17 |   24 |   59 |   22 |
        ae5-2.akbr7.global-gateway.net.nz -    0 |   51 |   51 |   17 |   30 |   55 |   24 |
     ae1-10.tkbr12.global-gateway.net.nz -    0 |   51 |   51 |   17 |   24 |   58 |   18 |
     xe0-1-0.sebr1.global-gateway.net.nz -    0 |   51 |   51 |   40 |   54 |   82 |   53 |
      ae1-10.sebr2.global-gateway.net.nz -    0 |   51 |   51 |   40 |   54 |   85 |   51 |
xe-3-1-0.a00.sydnau02.au.ra.gin.ntt.net -    0 |   51 |   51 |   40 |   54 |   81 |   48 |
xe-3-1-0.r00.sydnau02.au.bb.gin.ntt.net -    0 |   51 |   51 |  183 |  196 |  259 |  188 |
        as-0.r22.tokyjp01.jp.bb.gin.ntt.net -    0 |   51 |   51 |  196 |  217 |  253 |  210 |
       ae-5.r21.sngpsi05.sg.bb.gin.ntt.net -    0 |   51 |   51 |  261 |  299 |  359 |  266 |
     ae-13.r00.sngpsi02.sg.bb.gin.ntt.net -    0 |   51 |   51 |  298 |  314 |  342 |  301 |
                                   116.51.27.190 -    0 |   51 |   51 |  270 |  284 |  315 |  280 |
                                103.253.144.242 -    0 |   51 |   51 |  294 |  307 |  342 |  305 |
              do-sgp1.relay.mattcorallo.com -    0 |   51 |   51 |  272 |  284 |  316 |  282 |


Next the public.us-west.relay.mattcorallo.com which I'm using.


                                   fw                 -    0 |   52 |   52 |    0 |    0 |    1 |    0 |
          xxxxxx.jetstream.xtra.co.nz          -   0 |   52 |   52 |   14 |   24 |  40 |   26 |
mdr-ip20e-int.msc.global-gateway.net.nz -   0 |   52 |   52 |   18 |   25 |   41 |   26 |
       ae0-10.akbr6.global-gateway.net.nz -   0 |   52 |   52 |   17 |   25 |   37 |   29 |
        ae5-2.akbr7.global-gateway.net.nz -    0 |   52 |   52 |   17 |   25 |   41 |   29 |
     ae1-10.tkbr12.global-gateway.net.nz -    0 |   52 |   52 |   18 |   25 |   37 |   27 |
 xe7-0-1-10.lebr7.global-gateway.net.nz -    0 |   52 |   52 |  141 |  150 |  184 |  144 |
       ae3-10.sjbr3.global-gateway.net.nz -    0 |   52 |   52 |  145 |  162 |  188 |  154 |
           ae0.pabr5.global-gateway.net.nz -    0 |   52 |   52 |  145 |  152 |  169 |  146 |
        ae-8.r06.plalca01.us.bb.gin.ntt.net -    0 |   52 |   52 |  181 |  188 |  207 |  188 |
      ae-15.r02.snjsca04.us.bb.gin.ntt.net -    0 |   52 |   52 |  166 |  174 |  189 |  169 |
       ae-7.r21.snjsca04.us.bb.gin.ntt.net -    0 |   52 |   52 |  174 |  187 |  263 |  176 |
       ae-3.r20.sttlwa01.us.bb.gin.ntt.net -    3 |   48 |   47 |  193 |  196 |  221 |  194 |
       ae-1.r04.sttlwa01.us.bb.gin.ntt.net -    0 |   52 |   52 |  167 |  173 |  196 |  171 |
                       No response from host -  100 |   11 |    0 |    0 |    0 |    0 |    0 |
                       No response from host -  100 |   11 |    0 |    0 |    0 |    0 |    0 |
              rn-sea1.relay.mattcorallo.com -    0 |   52 |   52 |  193 |  200 |  219 |  194 |

:-)

[bryonp]

but hey guys what's going on? in the last 20 days p2p lost about 70 out of 250 active users / addesses (although the hashrate scores 1ph more), share diff of the pool is very high for those who do not have a farm, I personally with 3 th / s i have not done a good share in the last 20 hours ... there is no way of remodulating the share diff of p2p so that it is not so difficult for small-scale miners have a decent payout and a little more stable? ok accept the variance I can tolerate, but it seems clear that many users they are leaving the pool

E

All I can tell you is that I am pumping out 4TH and a few months ago I was getting around 100.oo a day give or take a few good days and a few bad days...BUT - now I am lucky to get 10.oo to 12.oo per block and with 2 to 4 per day I am making no money and my electric is way over 1000.00 per month.
I am just about to give up on p2pool and see what happens else where........ Its just not doing it for me these days?
Any thoughts from anyone is welcome .........
I love it here but I am just at a point of killing myself!
I hate to stop but its about that level I think....
Dont want to put another 5000.00 into machines to just break even in 3 months.....

Have you thought about the pencil mod for your S1s? You could still be in the game for a bit longer then sell them in a few more months. Just a thought.

Not really sure what the pencil does??? I did read it a few months ago but not sure I understood.... Any info? Thanks

[Duce]

but hey guys what's going on? in the last 20 days p2p lost about 70 out of 250 active users / addesses (although the hashrate scores 1ph more), share diff of the pool is very high for those who do not have a farm, I personally with 3 th / s i have not done a good share in the last 20 hours ... there is no way of remodulating the share diff of p2p so that it is not so difficult for small-scale miners have a decent payout and a little more stable? ok accept the variance I can tolerate, but it seems clear that many users they are leaving the pool

E

All I can tell you is that I am pumping out 4TH and a few months ago I was getting around 100.oo a day give or take a few good days and a few bad days...BUT - now I am lucky to get 10.oo to 12.oo per block and with 2 to 4 per day I am making no money and my electric is way over 1000.00 per month.
I am just about to give up on p2pool and see what happens else where........ Its just not doing it for me these days?
Any thoughts from anyone is welcome .........
I love it here but I am just at a point of killing myself!
I hate to stop but its about that level I think....
Dont want to put another 5000.00 into machines to just break even in 3 months.....

Have you thought about the pencil mod for your S1s? You could still be in the game for a bit longer then sell them in a few more months. Just a thought.

Not really sure what the pencil does??? I did read it a few months ago but not sure I understood.... Any info? Thanks

I ran mine like this as others with very good results in both power and stability. Here is the thread with some good guides. There is a resistor change you can perform as well but the quick and very easy approach is the pencil mod IMHO. https://bitcointalk.org/index.php?topic=526060.0

[Matt Corallo]

New Zealand

A Sydney or Auckland node would help the AUS/NZ region for sure.

You can see the routes below

...........

That is some strange routing indeed...NZ->JPY->SIN and NZ->SJC->SEA...are there any providers you know of in SIN that have proper routes from NZ? I was gonna move the SIN server to another provider in SIN anyway...

[windpath]

New Zealand

A Sydney or Auckland node would help the AUS/NZ region for sure.

You can see the routes below

...........

That is some strange routing indeed...NZ->JPY->SIN and NZ->SJC->SEA...are there any providers you know of in SIN that have proper routes from NZ? I was gonna move the SIN server to another provider in SIN anyway...

We have been experimenting with LBR (Latency Based Routing) on the AWS platform. The routes from "outliers" are almost always geographically bizarre, and as far as latency is concerned are about 90% accurate on the current fastest route. What I have realized is that they will pick the most reliable fastest route, which is not always fastest, but is more likely to reach its destination...

[windpath]

24 Hour average hash rate passes 3 PH/s!

[linuxforyou]

More and more blocks with p2pool which is nice, however Is it really useful now for smaller miners?

[coinme.info]

New Zealand

A Sydney or Auckland node would help the AUS/NZ region for sure.

You can see the routes below

...........

That is some strange routing indeed...NZ->JPY->SIN and NZ->SJC->SEA...are there any providers you know of in SIN that have proper routes from NZ? I was gonna move the SIN server to another provider in SIN anyway...

The best routes I've found for AUS/NZ are to the US on the Southern Cross cable.



The C - D - E1 route is

C -    ae1-10.tkbr12.global-gateway.net.nz -    0 |   52 |   52 |   18 |   25 |   37 |   27 |
    xe7-0-1-10.lebr7.global-gateway.net.nz -    0 |   52 |   52 |  141 |  150 |  184 |  144 |
E1 -    ae3-10.sjbr3.global-gateway.net.nz -    0 |   52 |   52 |  145 |  162 |  188 |  154 |

Not too bad for 12,000+ kms

Asia is all over the show as you can see with that SYD-JPY-SIN path using NTT.

Pacnet is a long time provider in the Asia Pacific area and have a new DC in SIN that may be an option

http://www.pacnet.com/data-center-services/hardware-as-a-service/

Not sure of the pricing for this, you'd have to drop them an email.

 

[EErik]

I totally agree with you and understand what is going on. I was just curious to hear from others. I understand the difficulty levels and the drop in pay.... As we all know, it dosent stop the electric bill and its HUGE.
I love mining and all of it, but I dont want to take my money and invest thousands more and be behind the eight ball for months to catch up like I did with this 5000.00 investment on the 4 T. I am finally ahead of the game and made a few bucks.
I am at that crossroad.... yes I could invest again and start all over...... I dont know its a dilemma that I have to come to grips with. Maybe I will sell off all of my S-1's and just run 1 S-2 at some pool to keep my hands and head in the game............  Thanks for your thoughts as thats all I was looking for!......
Nobody understands this crazy game we play but those of us wrapped up in it!!!

Couldn't agree more...

It's good that you've at least gotten to the point where you're a little ahead.  Sadly there's many who haven't and may never, and it'll be a completely losing proposition for them.  If you sell your equipment now you'll even be that much more ahead of the game, and like you said if you can get maybe just one or two things to keep it going as a hobby it'll be enough.  I think honestly that's how many will need to treat it - a hobby where they understand they may lose more than they invest but it's still kinda a neat thing to be part of.

Mining is certainly more about the passion and the geeky side of it than any rational explanation can bear.  I suffer the same disease...

if the share diff to raise the reward was lower, it would be much easier for small-scale miners have a constant reward, let us realize that now with 3ph/s is really shocking make a share and many are coming out.

The latest shares I made have given me 5mbtc which is very much higher than the daily reward of a small miner, the scale of the reward should be much more fragmented imho and, in this way, many more people participate

I do not understand the discourse that many are coming out of the pool because they have old hardware. there should be a replacement ... I think if people with older hardware comes out other with new should arrive, but it is not.

E

[coinme.info]

My node running Matt's fork with the embedded python-based relay code has been up for almost 24 hours now.  Initial reports seem that it's doing very well in general - over 110% efficiency for long periods, currently at 115%.  19 shares so far, all good - 0/0 orph/dead.

So, at the very least it's not hurting it at all, and hopefully is helping.  Will need to run longer to see how the share rates and efficiency do... I seem to see that as it goes along, getting past the 40 share mark is when I start seeing orphans/dead pop up.

On a side note tho - anyone seeing their DOA rates climb higher over the past week or so?  I was happily running around 3-4% DOA for the longest time but now I'm seeing it in the 6-8% range quite a lot.  Still better than the pool average, but I'm wondering if there's a core reason for the rates rising.  (and yes I've changed my S3's down from the default queue 4096, and I've tried 0 and 1 with neither being any majorly different).

Similar results to you using Matt's fork and the embedded python-based relay since I got this going on Windows Server 2012 R2 with an update from Matt.

Version: unknown 7032706f6f6c2d6d6173746572
Pool rate: 3.00PH/s (13% DOA+orphan) Share difficulty: 12700000
Node uptime: 14.9 hours Peers: 6 out, 3 in
Local rate: 2.38TH/s (3.6% DOA) Expected time to share: 6.4 hours
Shares: 2 total (0 orphaned, 0 dead) Efficiency: 115.0%

Not seeing the increase in DOA yet but I've probably jinxed it now so time will tell.

GBT latency is the big issue for me as I'm 12,000+km of fiber away from Matt's US-West node which I'm using as it was 100ms faster than the AU node.
This is normal as any US-West servers are normally lower latency than Asia nodes due to network topology.

Currently GBT is between 0.2 - 0.4 so I'm happy with that and the low DOA and normal shares being found seems to support that things are running ok.

Just a followup on Matt's fork with the python-based relay included.

Version: unknown 7032706f6f6c2d6d6173746572
Pool rate: 2.45PH/s (13% DOA+orphan) Share difficulty: 11500000
Node uptime: 13.5 hours Peers: 6 out, 5 in
Local rate: 3.03TH/s (1.7% DOA) Expected time to share: 4.5 hours
Shares: 2 total (0 orphaned, 0 dead) Efficiency: 115.0%

Did a restart of P2pool unfortunately as something maxed out my upload bandwidth otherwise would have a full 24+ hour snap shot.
Turned out the be a web browser client & not p2pool but now I know to trust p2pool over any other app!

So low DOA and a respectable Getwork Latency Mean of 0.305s using Antminer S1s and S3s

Matt's relay node mod has made a real difference to my p2pool node.

Thanks Matt!

[jedimstr]

Nice results on the RelayNode mod... But how does it compare to having the standalone java node and p2pool instance versus the combined python implementation?  Pros/cons? Would pypy execution make any appreciable difference for the modified p2pool instance? 

Right now I'm running the mainline p2pool via pypy and Matt's java Relay separately and it's functioning fine, so I'm interested if there would be any gain to switch to the combined implementation.

[mdude77]

Nice results on the RelayNode mod... But how does it compare to having the standalone java node and p2pool instance versus the combined python implementation?  Pros/cons? Would pypy execution make any appreciable difference for the modified p2pool instance? 

Right now I'm running the mainline p2pool via pypy and Matt's java Relay separately and it's functioning fine, so I'm interested if there would be any gain to switch to the combined implementation.

The biggest gain I'd say to using the combined version is that you won't need Java on your machine.

I'm going to switch my public node to using the combined version, see if it helps.

M

[Duce]

Nice results on the RelayNode mod... But how does it compare to having the standalone java node and p2pool instance versus the combined python implementation?  Pros/cons? Would pypy execution make any appreciable difference for the modified p2pool instance? 

Right now I'm running the mainline p2pool via pypy and Matt's java Relay separately and it's functioning fine, so I'm interested if there would be any gain to switch to the combined implementation.

The biggest gain I'd say to using the combined version is that you won't need Java on your machine.

I'm going to switch my public node to using the combined version, see if it helps.

M

I see that you changed your pool to Matt's fork (python included) as well. Did you try to use your extended node status and found that it did not work? I too have seen an improvement as others with Matt's fork but lost the ability to run a parallel version. I am really asking to verify that it is not just me? This is by no means a discredit to Matt's efforts as the loss of a "pretty" display does not take away from the improved performance.

[mdude77]

Nice results on the RelayNode mod... But how does it compare to having the standalone java node and p2pool instance versus the combined python implementation?  Pros/cons? Would pypy execution make any appreciable difference for the modified p2pool instance? 

Right now I'm running the mainline p2pool via pypy and Matt's java Relay separately and it's functioning fine, so I'm interested if there would be any gain to switch to the combined implementation.

The biggest gain I'd say to using the combined version is that you won't need Java on your machine.

I'm going to switch my public node to using the combined version, see if it helps.

M

I see that you changed your pool to Matt's fork (python included) as well. Did you try to use your extended node status and found that it did not work? I too have seen an improvement as others with Matt's fork but lost the ability to run a parallel version. I am really asking to verify that it is not just me? This is by no means a discredit to Matt's efforts as the loss of a "pretty" display does not take away from the improved performance.

Not sure what you mean by extended node status?

I still have a pretty display, as far as I can tell.  Not the best, but better than the default.

M

[Duce]

Nice results on the RelayNode mod... But how does it compare to having the standalone java node and p2pool instance versus the combined python implementation?  Pros/cons? Would pypy execution make any appreciable difference for the modified p2pool instance? 

Right now I'm running the mainline p2pool via pypy and Matt's java Relay separately and it's functioning fine, so I'm interested if there would be any gain to switch to the combined implementation.

The biggest gain I'd say to using the combined version is that you won't need Java on your machine.

I'm going to switch my public node to using the combined version, see if it helps.

M

I see that you changed your pool to Matt's fork (python included) as well. Did you try to use your extended node status and found that it did not work? I too have seen an improvement as others with Matt's fork but lost the ability to run a parallel version. I am really asking to verify that it is not just me? This is by no means a discredit to Matt's efforts as the loss of a "pretty" display does not take away from the improved performance.

Not sure what you mean by extended node status?

I still have a pretty display, as far as I can tell.  Not the best, but better than the default.

M

I can't get anything but the default to work, I built the node twice to make sure I did not do something wrong but it does not seem to pull the data in. I must be missing something. When I access the non-default front end it shows up on the web but does not display any data, it worked before this build. Well I would rather let it run than waste the time to playing with another front end.

[mdude77]

Nice results on the RelayNode mod... But how does it compare to having the standalone java node and p2pool instance versus the combined python implementation?  Pros/cons? Would pypy execution make any appreciable difference for the modified p2pool instance? 

Right now I'm running the mainline p2pool via pypy and Matt's java Relay separately and it's functioning fine, so I'm interested if there would be any gain to switch to the combined implementation.

The biggest gain I'd say to using the combined version is that you won't need Java on your machine.

I'm going to switch my public node to using the combined version, see if it helps.

M

I see that you changed your pool to Matt's fork (python included) as well. Did you try to use your extended node status and found that it did not work? I too have seen an improvement as others with Matt's fork but lost the ability to run a parallel version. I am really asking to verify that it is not just me? This is by no means a discredit to Matt's efforts as the loss of a "pretty" display does not take away from the improved performance.

Not sure what you mean by extended node status?

I still have a pretty display, as far as I can tell.  Not the best, but better than the default.

M

I can't get anything but the default to work, I built the node twice to make sure I did not do something wrong but it does not seem to pull the data in. I must be missing something. When I access the non-default front end it shows up on the web but does not display any data, it worked before this build. Well I would rather let it run than waste the time to playing with another front end.

All I did was:

- rename my existing p2pool folder
- git clone the fork
- move the data folder from the old p2pool folder to the new one
- move/copy the webstatic folder from the old p2pool folder to the new one

And of course modified my startup script for p2pool to use the relay code.

M

[PatMan]

All I did was:

- rename my existing p2pool folder
- git clone the fork
- move the data folder from the old p2pool folder to the new one
- move/copy the webstatic folder from the old p2pool folder to the new one

And of course modified my startup script for p2pool to use the relay code.

M

Same here, no problems at all.

[Duce]

All I did was:

- rename my existing p2pool folder
- git clone the fork
- move the data folder from the old p2pool folder to the new one
- move/copy the webstatic folder from the old p2pool folder to the new one

And of course modified my startup script for p2pool to use the relay code.

M

Same here, no problems at all.

Wow thanks guys, I didn't think about just copying the old web-static folder over but relied on just making a new one. I did rename the existing p2pool folder so I do have the old files. That was easy, thanks again!

[hamburgerhelper]

Nice to see Matt being so responsive to the folks who are Java-averse. That's the way you build a user base!

Has anyone who ran p2pool + RelayNodeClient.jar compared the resource utilization with the 100% python version (Matt's fork of p2pool)? I'm interested in either a decrease in RAM or CPU utilization. I've been running the Java client on 2GB of RAM w/2 CPUs and haven't noticed trouble. I'm looking for some hard data to convince me that the switch to pure python is worth it.

Thanks for helping me be lazy and still run a bitchin p2pool node.

[Matt Corallo]

Nice results on the RelayNode mod... But how does it compare to having the standalone java node and p2pool instance versus the combined python implementation?  Pros/cons? Would pypy execution make any appreciable difference for the modified p2pool instance? 

Right now I'm running the mainline p2pool via pypy and Matt's java Relay separately and it's functioning fine, so I'm interested if there would be any gain to switch to the combined implementation.

The only thing I can think of is that, because Python is single-threaded in practice (both pypy and CPython have this global lock (the GIL) and they run multiple threads which contend for it, actually decreasing performance over just running a single thread and multiplexing between them....), you may end up in a situation where you're waiting for the relay network thread to finish processing something, but because all of its processing is incredibly light-weight, I highly doubt it would make more than a few nanoseconds difference.