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

[gnar1ta$]

Yeah I like how everyone sees the hashrate on cgminer and gets excited.  It doesn't really matter. cgminer reports over 530 G for me on btcguild, but the pool only reports 480-490 G.  Maybe I can get Knc to pay the difference since they refuse to prove they didn't make cgminer lie.  At least p2pool pays me more than cgminer is reporting 

[matthewh3]

Yeah I like how everyone sees the hashrate on cgminer and gets excited.  It doesn't really matter. cgminer reports over 530 G for me on btcguild, but the pool only reports 480-490 G.  Maybe I can get Knc to pay the difference since they refuse to prove they didn't make cgminer lie.  At least p2pool pays me more than cgminer is reporting 

I had a solid 560GH reported by Slush over ten rounds before the board died.  It's due to arrive in Sweden tomorrow and only then can they replace it.  160GH is a lot too loose and I want compo but was planning BabyJets next if they ship this month.

[gnar1ta$]

Yeah I like how everyone sees the hashrate on cgminer and gets excited.  It doesn't really matter. cgminer reports over 530 G for me on btcguild, but the pool only reports 480-490 G.  Maybe I can get Knc to pay the difference since they refuse to prove they didn't make cgminer lie.  At least p2pool pays me more than cgminer is reporting 

I had a solid 560GH reported by Slush over ten rounds before the board died.  It's due to arrive in Sweden tomorrow and only then can they replace it.  160GH is a lot too loose and I want compo but was planning BabyJets next if they ship this month.

Hmmm....maybe I just have an extra crappy model.  I've tried 3 different pools and get 50-80G difference from what the pool reports and what cgminer reports.  Though they all report much higher than p2pool.

[matthewh3]

Yeah I like how everyone sees the hashrate on cgminer and gets excited.  It doesn't really matter. cgminer reports over 530 G for me on btcguild, but the pool only reports 480-490 G.  Maybe I can get Knc to pay the difference since they refuse to prove they didn't make cgminer lie.  At least p2pool pays me more than cgminer is reporting 

I had a solid 560GH reported by Slush over ten rounds before the board died.  It's due to arrive in Sweden tomorrow and only then can they replace it.  160GH is a lot too loose and I want compo but was planning BabyJets next if they ship this month.

Hmmm....maybe I just have an extra crappy model.  I've tried 3 different pools and get 50-80G difference from what the pool reports and what cgminer reports.  Though they all report much higher than p2pool.

I've currently got 398959MH off just three chips over ten rounds on Slush.  With the fourth board and chip due at KNC in Sweden tomorrow for replacement.

[kano]

Yeah I like how everyone sees the hashrate on cgminer and gets excited.  It doesn't really matter. cgminer reports over 530 G for me on btcguild, but the pool only reports 480-490 G.  Maybe I can get Knc to pay the difference since they refuse to prove they didn't make cgminer lie.  At least p2pool pays me more than cgminer is reporting 

I had a solid 560GH reported by Slush over ten rounds before the board died.  It's due to arrive in Sweden tomorrow and only then can they replace it.  160GH is a lot too loose and I want compo but was planning BabyJets next if they ship this month.

Post a screen shot
... as I did above, it's easy to work out what it is really doing.

[AtomSea]

Just wanted to say thanks to everyone here.
We have a strong & free Litecoin and Bitcoin mining site because of your work.
Very, very appreciative.

[matthewh3]

Yeah I like how everyone sees the hashrate on cgminer and gets excited.  It doesn't really matter. cgminer reports over 530 G for me on btcguild, but the pool only reports 480-490 G.  Maybe I can get Knc to pay the difference since they refuse to prove they didn't make cgminer lie.  At least p2pool pays me more than cgminer is reporting  

I had a solid 560GH reported by Slush over ten rounds before the board died.  It's due to arrive in Sweden tomorrow and only then can they replace it.  160GH is a lot too loose and I want compo but was planning BabyJets next if they ship this month.

Post a screen shot
... as I did above, it's easy to work out what it is really doing.

I can only post one of the pool as my associate state side is hosting my Jupiter's for import VAT reasons here in the UK.

[zvs]

” Graphs

Version: 13.3

Pool rate: 15.9TH/s (12% DOA+orphan) Share difficulty: 70300

Node uptime: 39.446 days Peers: 6 out, 0 in

Local rate: 389GH/s (14% DOA) Expected time to share: 0.216 hours

Shares: 1246 total (183 orphaned, 184 dead) Efficiency: 80.54% “


my speed is  82g,but in p2pool.info it show  my speed is  66g !

does everyone knows about my speed is slower?

why it show :  Efficiency: 80.54%  and  66g ?

that pool has an excessive # of orphans, if you're only going to have 6 outgoing connections and 0 incoming, you should use addnode (ed: --p2pool-node) to make sure you get at least a couple that are good

use http://p2pool-nodes.info/  and look for some that are well connected and allow incoming connections, you should also add that 4.5TH one regardless, if it still allows incoming connections, not sure

[oldbushie]

I'm struggling with latency... for the first 15 hours when I started up p2pool my latency was 0.2 seconds. Then it inexplicably jumped to 1s at 7:30am a couple days ago (I was asleep but my dad turned on his Mac Mini apparently). I'm running it on an idle laptop that's not currently being used for anything else. I've tried:
- Turning off all other internet connected devices in the house to see if one could be the cause
- using QoS to prioritize bitcoind and p2pool packets
- fiddling with max connections, either low (20) or high (80) on both bitcoind and p2pool
- resetting the laptop p2pool is hosted on

I'm just boggling at what could possibly be the cause. The only things left I haven't tried are disabling/tweaking Kaspersky (any suggestions? not quite sure how to fine tune that besides having it exclude those directories from scanning, tho I do have Kaspersky Internet Security and it may be impeding the internal internet connections) or rebooting the router, maybe. I'm just not clear what suddenly triggered the latency to go into overdrive. Is bitcoind somehow remembering a bad node? How do I check for and/or clear bad nodes?

[oldbushie]

I'm trying something different... just turning off all extraneous settings except the ones needed barebones to make the setup work at all. Instead of having a few options turned on that I know nothing about. Let's see how long it stays reasonable.

[abbeytim]

is the laptop on wifi ?? if so maybe the latency is from the wifi connection

[zvs]

I'm struggling with latency... for the first 15 hours when I started up p2pool my latency was 0.2 seconds. Then it inexplicably jumped to 1s at 7:30am a couple days ago (I was asleep but my dad turned on his Mac Mini apparently). I'm running it on an idle laptop that's not currently being used for anything else. I've tried:
- Turning off all other internet connected devices in the house to see if one could be the cause
- using QoS to prioritize bitcoind and p2pool packets
- fiddling with max connections, either low (20) or high (80) on both bitcoind and p2pool
- resetting the laptop p2pool is hosted on

I'm just boggling at what could possibly be the cause. The only things left I haven't tried are disabling/tweaking Kaspersky (any suggestions? not quite sure how to fine tune that besides having it exclude those directories from scanning, tho I do have Kaspersky Internet Security and it may be impeding the internal internet connections) or rebooting the router, maybe. I'm just not clear what suddenly triggered the latency to go into overdrive. Is bitcoind somehow remembering a bad node? How do I check for and/or clear bad nodes?

wait, you mean the getblocktemplate latency is 1000ms?   remote stuff like google.com or something is 1000ms?  local network is 1000ms?

if it's #3, that should be easy to troubleshoot.  if it's #2, your upstream is probably becoming saturated (or your dad might be torrenting pr0n), so you should leave p2pool at 6 outbound connections and firewall 9332, change bitcoin.conf to listen=0, maxconnections=4, have enough connects (connect=xx.xx.xx.xx) so you dont get random connections (stuff i've been using for like a year, some like 5.9.24.81, 69.195.155.227, 188.40.112.72, 5.9.157.50, 89.238.64.139, 50.31.149.57).. some of those might not allow incoming connections anymore, i haven't maintenanced my addnode list in like 3 months..  but, uh, luke.dashjr.org/programs/bitcoin/files/charts/seeds.txt is a good reference (ignore any node that's using something older than 0.8.2, or maybe 0.8.1 is ok..  the bitpeer node doesnt do any transaction verification (or blocks), so that's nixed too)

if it's #1, you either need to update your bitcoind version (it's probably taking all those horse staple battery transactions), change the fees, or reduce maximum block size.  most likely the first

[oldbushie]

Yeah just changing everything to default settings fixed it... all I have set now on bitcoind is server mode and user/pass. I also got rid of the max-conns argument in my p2pool batch file.

[amer]

Can someone post a 24 hour chart of a BE blade's Hashrate+DOA rate? I want to see what "normal" is.

[rav3n_pl]

BE BLADE will have like 30% doa/orphans. It is NOT supporting LP so it is NOT good for P2pool.

[zvs]

I like how it shows the 'my shares' on the web interface now..

Any chance of getting it to retain local shares that are declared orphans?  i.e. I see today that one share was orphaned, but when I click on it in 'my shares' it's already been deleted by p2pool.  I'd like to see what orphaned it and why (was it the same person getting two shares in a row, is there some other node I need to connect to, just bad luck, or what)..

re: I used to get like 0-5% orphans, now it's more like 5-10%.     http://www.nogleg.com:9332/static/graphs.html?Year  vs  http://www.nogleg.com:9332/static/graphs.html?Week

[jedimstr]

Is there any way to manually add a node for outgoing without bringing my node down?  I tried running a parallel p2pool process after editing my configuration file with additional -n entries and  switching over by exiting the original process to minimize downtime (one of the suggested ways to minimize downtime when upgrading p2pool as p2pool can reconnect to the recently vacated ports) but this method seriously raised DOA count and killed efficiency more than just sequentially bringing down and restarting p2pool.

Hypothetical: would an optimized Multi-threaded C version of p2pool be a better future path, possibly reducing inefficiencies and latency versus the current Python version and allowing configuration changes like peer node additions and maintenance on the fly? 

[rav3n_pl]

You can run second instance, wait till it will cry about busy worker port then kill old instance.
You will lose like 10-15 seconds.

[jedimstr]

You can run second instance, wait till it will cry about busy worker port then kill old instance.
You will lose like 10-15 seconds.

Yeah, I mentioned I did that in my original post and mentioned that it killed efficiency for the node and raised DOA to a crazy number. That's why I was asking for any possible OTHER way of doing it.  Trying to do it this way gives worse results than just sequentially shutting down and restarting a single instance.

[rav3n_pl]

To avoid "large doa amount" you need shut down node for more than 1 minute to break all miners.
So kill, wait 30 sec, restart (will take like 1 minute), be happy of good stats.