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

[p3yot33at3r]

Code:

minrelaytxfee=0.0001
mintxfee=0.0001

What are the standard settings if nothing is specified in the .conf file? I read about changing these in a guide alse where, but it was a little dated so decided against it.

Edit: Aaaaand another block!! 

[1713555062]

1713555062

[1713555062]

1713555062

[1713555062]

1713555062

[1713555062]

1713555062

[Meuh6879]

I had an issue like the above on my two p2pool nodes as well. It appears to be due to bitcoind getting overloaded from the stress test. I added this to my .bitcoin/bitcoin.conf file:

Code:

minrelaytxfee=0.0001
mintxfee=0.0001

After that, performance improved and the periodic bitcoind dropouts stopped.

I don't know what the threshold is for getting it to work. It's possible that 0.00005 or lower might work as well. Anyone want to play around with different values? I've got over 350 TH/s on my nodes, so I don't want to do anything with them that would cause unnecessary downtime.

Aha ! (yes) ... that's why i have, now, a saturated Upload ...  

[jtoomim]

Code:

minrelaytxfee=0.0001
mintxfee=0.0001

What are the standard settings if nothing is specified in the .conf file? I read about changing these in a guide alse where, but it was a little dated so decided against it.

Edit: Aaaaand another block!! 

I believe the defaults are 0 -- i.e., bitcoind will fill blocks with zero-fee transactions if no fee-paying transactions are available, as long as those free transactions are large and old enough.

I will probably upgrade our main node's hardware soon so it can better handle these spammy transaction volumes. I'm trying to decide if I should wait for Skylake to be released (August, supposedly) or just fork out for an i7 4790k now. Currently, the faster of my nodes runs on a Core i3 3.3 GHz (dual-core) with 4 GiB of RAM. I have seen the RAM get saturated on these nodes before and revert to swapping. I've got an 8 GiB RAM upgrade kit that arrived yesterday, and that also might be enough to solve the problem.

[Meuh6879]

This day, "input already spend" overload my upload bandwidth ...
When i shutdown the P2Pool node, this isn't display (debug.log).

Have you somes tips to solve this ?
It's the first time that it's happend ...  

Bandwidth on P2Pool graphic is, it's true, saturated.

[windpath]

Bitcoin Core 0.11.0 released

I will not have time to test it today, but it looks like there are a lot of speed improvements for the mempool that should speed up GBT latency.

https://bitcoin.org/en/release/v0.11.0

Would be interested in hearing anybody's experience

[Meuh6879]

i use the RC2-RC3 since 2 weeks ...   especially to test the pruning mode at the beginning.

[squidicuz]

Bitcoin Core 0.11.0 released

I will not have time to test it today, but it looks like there are a lot of speed improvements for the mempool that should speed up GBT latency.

https://bitcoin.org/en/release/v0.11.0

Would be interested in hearing anybody's experience

I have been using 0.11 for a couple weeks now.  It works with latest p2pool, no issues, compiled from a local git clone.   Working with the current version of the relaynodeclient as well.

This day, "input already spend" overload my upload bandwidth ...
When i shutdown the P2Pool node, this isn't display (debug.log).

Have you somes tips to solve this ?
It's the first time that it's happend ...  

Bandwidth on P2Pool graphic is, it's true, saturated.

http://imagize[Suspicious link removed]ageshack.us/a/img633/7669/GKPWhg.jpg

If bandwidth is a major issue, you can try lowering the maximum allowed connected nodes.  I personally have not seen what you are experiencing though...

Code:

--max-conns CONNS

[luthermarcus]

Anyone know how to ban peers on bitcoin qt. I'm not sure it's possible.
Also can anyone point me in direction of a good tutorial for bitcoind?

Bitcoin 0.11.0 solves the freezing i was seeing (recommended to upgrade if you have issues)  This release supports running a fully validating node without maintaining a copy of the raw block and undo data on disk. AKA pruning which has to be manually activated by using below code in
bitcoin config and don't use it if you use the wallet function because it turns it off. (also incompatible with -txindex will auto. turn it of)

prune=N : where N is the number of MB to allot for raw block & undo data (don't know what is a good number they mention 550mb though)

They are also recommending these for the transaction flooding

minrelaytxfee=0.00005
limitfreerelay=5  defualt=10

[Meuh6879]

If bandwidth is a major issue, you can try lowering the maximum allowed connected nodes.  I personally have not seen what you are experiencing though...

Code:

--max-conns CONNS

  thanks for this tips.

problem solve ... the shares of P2Pool are corrupted (i delete it, and after this, no problem with bandwidth).
strange.
first time.

[jonnybravo0311]

Anyone know how to ban peers on bitcoin qt. I'm not sure it's possible.
Also can anyone point me in direction of a good tutorial for bitcoind?

Bitcoin 0.11.0 solves the freezing i was seeing (recommended to upgrade if you have issues)  This release supports running a fully validating node without maintaining a copy of the raw block and undo data on disk. AKA pruning which has to be manually activated by using below code in
bitcoin config and don't use it if you use the wallet function because it turns it off. (also incompatible with -txindex will auto. turn it of)

prune=N : where N is the number of MB to allot for raw block & undo data (don't know what is a good number they mention 550mb though)

They are also recommending these for the transaction flooding

minrelaytxfee=0.00005
limitfreerelay=5  defualt=10

I'm not sure you can ban peers directly with the core software.  If needed, I just add firewall rules to block the IP addresses of the bad peers.  If you're using Ubuntu and ufw, it's absurdly easy:

Code:

ufw deny from IP_ADDRESS

My local node has been updated to core 0.11.0 and 14.0-9.  While I'm still seeing higher-than-usual rejected hash rate percentages, I haven't experienced the freezing/lost contact issues again.  So far, so good.

[Automatic Monkey]

Hey, can anybody recommend some good nodes to use for Nicehash (Amsterdam) and Westhash (San Jose)? I'm ready to drop some rental hash on it, when the price is right.

[jonnybravo0311]

OgNasty runs his node in the EU (Germany): https://nastyfans.org/nastypool/.  Not sure about west coast US nodes, though.

[luthermarcus]

Anyone know how to ban peers on bitcoin qt. I'm not sure it's possible.
Also can anyone point me in direction of a good tutorial for bitcoind?

Bitcoin 0.11.0 solves the freezing i was seeing (recommended to upgrade if you have issues)  This release supports running a fully validating node without maintaining a copy of the raw block and undo data on disk. AKA pruning which has to be manually activated by using below code in
bitcoin config and don't use it if you use the wallet function because it turns it off. (also incompatible with -txindex will auto. turn it of)

prune=N : where N is the number of MB to allot for raw block & undo data (don't know what is a good number they mention 550mb though)

They are also recommending these for the transaction flooding

minrelaytxfee=0.00005
limitfreerelay=5  defualt=10

I'm not sure you can ban peers directly with the core software.  If needed, I just add firewall rules to block the IP addresses of the bad peers.  If you're using Ubuntu and ufw, it's absurdly easy:

Code:

ufw deny from IP_ADDRESS

My local node has been updated to core 0.11.0 and 14.0-9.  While I'm still seeing higher-than-usual rejected hash rate percentages, I haven't experienced the freezing/lost contact issues again.  So far, so good.

I heard blocking peers can effect your bitcoin node in a weird way. If you are using    -addnode or -connect try using defualts and see if that fixes it.Depending on your peers take off firewall as well.

[p3yot33at3r]

I heard blocking peers can effect your bitcoin node in a weird way. If you are using    -addnode or -connect try using defualts and see if that fixes it.Depending on your peers take off firewall as well.

Where did you hear this - can you provide a link? I've never had any issues.....

Bitcoin Core 0.11.0 released

I will not have time to test it today, but it looks like there are a lot of speed improvements for the mempool that should speed up GBT latency.

https://bitcoin.org/en/release/v0.11.0

Would be interested in hearing anybody's experience

Been running this version since it was released, so far it's proven very stable with no disconnects - although I do have quite a few bad peers blocked, weather this has something to do with it I'm not sure, but generally I'm very happy with how it performs.

EDIT: Although I am having trouble with the relaynetworkclient disconnecting - anyone else having this problem?

[Meuh6879]

yes, mainly because of a corrupted share chain (of P2Pool).
verified share can't rease, too ... in this case of many disconnected clients.

i must have deleted 3 times (in 1 week) the sharechain of P2Pool to have a standard P2Pool reaction.

probably the old v13 that it tries to inject shares ... again.

[p3yot33at3r]

yes, mainly because of a corrupted share chain (of P2Pool).
verified share can't rease, too ... in this case of many disconnected clients.

i must have deleted 3 times (in 1 week) the sharechain of P2Pool to have a standard P2Pool reaction.

probably the old v13 that it tries to inject shares ... again.

I too experienced a corrupted p2pool sharechain & had to delete/redownload a new one a few days ago. It seems like p2pool is not fully banning older v13 releases, which seems to be causing the problem. Since I started proactively banning bad peers manually the problems have all but gone, but having to constantly monitor my logs for repeat offending IP's is a real bind & very laborious.
Also, looking around the various p2pool node finder sites, there's still a multitude of v13 nodes still mining away regardless - which makes me wonder if these nodes are actually still getting paid, as there have been zero complaints here about unpaid blocks. If these outdated nodes are still getting paid, that would mean that their owners are still blissfully unaware of any required upgrade & would also explain why there are still so many older nodes humming away......

[iSOcH]

Hey, can anybody recommend some good nodes to use for Nicehash (Amsterdam) and Westhash (San Jose)? I'm ready to drop some rental hash on it, when the price is right.

i'm also running a node in germany; https://pascal-schwarz.ch/p2pool-node-status/

[idonothave]

Howto get latency of bitcoin-qt low? I am just now around 0.5s

Install bitcoin v0.11. Add to your bitcoin.conf:

maxconnections=10

then run bitcoin-qt and look connections in debug witdow. Choose IPs with lowest latencies and add them to your bitcoin.conf. Add there not more then 4 the best.

addnode=x.x.x.x

[luthermarcus]

yes, mainly because of a corrupted share chain (of P2Pool).
verified share can't rease, too ... in this case of many disconnected clients.

i must have deleted 3 times (in 1 week) the sharechain of P2Pool to have a standard P2Pool reaction.

probably the old v13 that it tries to inject shares ... again.

I too experienced a corrupted p2pool sharechain & had to delete/redownload a new one a few days ago. It seems like p2pool is not fully banning older v13 releases, which seems to be causing the problem. Since I started proactively banning bad peers manually the problems have all but gone, but having to constantly monitor my logs for repeat offending IP's is a real bind & very laborious.
Also, looking around the various p2pool node finder sites, there's still a multitude of v13 nodes still mining away regardless - which makes me wonder if these nodes are actually still getting paid, as there have been zero complaints here about unpaid blocks. If these outdated nodes are still getting paid, that would mean that their owners are still blissfully unaware of any required upgrade & would also explain why there are still so many older nodes humming away......

How do you figure your sharechain was corrupted?
Mine seems to be kicking peers very often. Could that traffic just be rerouted to a dead zone where we dont have to worry about them affecting the majority who actually try to keep things up to date. It also ends up sending them a share before it kicks them could that be slowing the network down.

Just in case someone wants to block these peers manually i do this on windows.
https://serverfault.com/questions/117420/windows-firewall-ip-ban

[p3yot33at3r]

How do you figure your sharechain was corrupted?

I'm not sure, I can only put it down to the bad peers, which I've now blocked - no problems since