Author

Topic: [1500 TH] p2pool: Decentralized, DoS-resistant, Hop-Proof pool - page 765. (Read 2591920 times)

sr. member
Activity: 409
Merit: 251
Crypt'n Since 2011
Is it possible to turn off dynamic share difficulty for local shares?

Quote
New work for workers! Difficulty: 9.8273 Share Difficulty 635.184144

I don't really see the point of this.  Even at 10GH/s locally the amount of miner traffic is minimal and it is on the LAN.  Keeping shares difficult 1 locally allows apples to apples comparison at the miner level.

With dynamic difficulty changing I just realized it made comparing stats (in cgminer) between rigs with different uptimes impossible.  cgminer reports total # of shares but those shares represent different amounts of work.    I was comparing mem clock difference between two rigs and the stats are junk.  Stupid me for not thinking of this before running the trial.

Is there a command line argument to specify a static local share value?  If not can it be included in the next version?



If you comment out line 533-538 in the main.py file (newest version) it will serve only diffficulty 1 shares to your miner

See before and after below.

Before:
Code:
                target = net.PARENT.SANE_MAX_TARGET
                if len(self.recent_shares_ts_work) == 50:
                    hash_rate = sum(work for ts, work in self.recent_shares_ts_work)//(self.recent_shares_ts_work[-1][0] - self.recent_shares_ts_work[0][0])
                    target = min(target, 2**256//(hash_rate))
                target = max(target, share_info['bits'].target)
                for aux_work in current_work.value['mm_chains'].itervalues():
                    target = max(target, aux_work['target'])

After:
Code:
                target = net.PARENT.SANE_MAX_TARGET
                # if len(self.recent_shares_ts_work) == 50:
                #  hash_rate = sum(work for ts, work in self.recent_shares_ts_work)//(self.recent_shares_ts_work[-1][0] - self.recent_shares_ts_work[0][0])
                #  target = min(target, 2**256//(hash_rate))
                #  target = max(target, share_info['bits'].target)
                #for aux_work in current_work.value['mm_chains'].itervalues():
                #    target = max(target, aux_work['target'])

Or you can run your miner at http://p2pmining.com

 Wink
donator
Activity: 1218
Merit: 1079
Gerald Davis
Is it possible to turn off dynamic share difficulty for local shares?

Quote
New work for workers! Difficulty: 9.8273 Share Difficulty 635.184144

I don't really see the point of this.  Even at 10GH/s locally the amount of miner traffic is minimal and it is on the LAN.  Keeping shares difficult 1 locally allows apples to apples comparison at the miner level.

With dynamic difficulty changing I just realized it made comparing stats (in cgminer) between rigs with different uptimes impossible.  cgminer reports total # of shares but those shares represent different amounts of work.    I was comparing mem clock difference between two rigs and the stats are junk.  Stupid me for not thinking of this before running the trial.

Is there a command line argument to specify a static local share value?  If not can it be included in the next version?

legendary
Activity: 1442
Merit: 1000
was this block: http://blockchain.info/tx/369970d60ba54bae122be472366938626d2533e2f79cdda407e48eaa3765c68a an orpan?

I never got a payout,but got one for the block after (been mining for weeks) Huh

That's a transaction, not a block; and no, the associated block wasn't orphaned. If you weren't paid, you either didn't submit a share the previous 24h or mined on a different p2pool fork.

Rebooted PC and the payment showed up in the client. Smiley
donator
Activity: 1218
Merit: 1079
Gerald Davis
Thanks for all your hard work at keeping Bitcoin decentralized and secure forrestv.

Donation sent:
5f6215e518678a76b70dc3a14e4275559f5fd05c5755e844f7575f1e856bd0fa
hero member
Activity: 516
Merit: 643
P2Pool release 0.9.1 - tag: 0.9.1 - UPGRADE TO >=0.9 REQUIRED before Mar 4, upgrade to 0.9.1 will make transition go more smoothly

Windows py2exe binary: http://u.forre.st/u/jkwntafq/p2pool_win32_1cf73bc.zip
Source tarball: https://github.com/forrestv/p2pool/tarball/0.9.1
Source zip: https://github.com/forrestv/p2pool/zipball/0.9.1

Changes:
* Some fixes to make the transition on March 4th go better
** P2Pool will switch using the sharechain's time instead of your local clock's time, so people with very incorrect clocks won't be hurt
** P2Pool will disconnect from and ban unupgraded peers so you won't be spammed with error messages from the few people running older versions


Persistent note: I would recommend switching to Bitcoin 0.6.0 RC 1, which includes the RPC getblock call. The getblock call lets P2Pool keep track of block heights more robustly, and so might protect you from sharechain forks. Download it from https://bitcointalksearch.org/topic/version-06-release-candidate-1-63165
full member
Activity: 373
Merit: 100
was this block: http://blockchain.info/tx/369970d60ba54bae122be472366938626d2533e2f79cdda407e48eaa3765c68a an orpan?

I never got a payout,but got one for the block after (been mining for weeks) Huh

That's a transaction, not a block; and no, the associated block wasn't orphaned. If you weren't paid, you either didn't submit a share the previous 24h or mined on a different p2pool fork.
legendary
Activity: 1442
Merit: 1000
was this block: http://blockchain.info/tx/369970d60ba54bae122be472366938626d2533e2f79cdda407e48eaa3765c68a an orpan?

I never got a payout,but got one for the block after (been mining for weeks) Huh
sr. member
Activity: 459
Merit: 250
My client is going to IRC and I do have a port open for incoming connections.

If you want to give it a try, connect to the ip of blockexplorer.funkymonkey.org port 8368.  (addnode=ip:port in your coiledcoin.conf file)

I've got a GroupCoin client running but no one to connect to so it's sitting at block 0.

.. we should probably start a separate thread for this so we don't derail this one and so others can see if interested... Smiley


edit:  I joined all 5 irc channels with my desktop client, the first channel 00 is the only one with users and you're one of them... not sure why you're not trying to connect..

edit 2:  I can open ports for groupcoin however, I wouldn't hard code my IP into anything.. Everything I have runs off a residential internet connection.  My IP hasn't changed in a couple of years but it does change from time to time and there's no guaranteed uptime here from my provider.  Wink
legendary
Activity: 2940
Merit: 1090
I am pretty sure the MiniUPNP lib does not work with my router, so I probably do not have incoming ports. However if you don't have IRC turned off presumably my client should see yours in the IRC channel and call out to it.

Maybe though they forgot to limit the number of IRC channels, leaving a vanishingly small chance that any two clients happen to randomly choose the same channel, or maybe some versions use one IRC network and others another.

...There are five channels, 00 through 04, so only a one in five chance we land in the same IRC channel.

There probably should be a commandline and/or config setting to fixate on a specific channel but I do not think there is.

Maybe also it should check the channel actually has clients in it and if not try another.

Maybe whoever runs the main nodes the fallback methods use has shut down or something, I wonder if they have a DNS thing set up for finding fallback nodes, or a list or something?

If you have an incoming port, I can tell mine specifically on the commandline to connect to yours, most of my shell scripts for startign coin daemons have a few nodes hardcoded into them.

While we're at it, if you can open incoming ports, maybe you could run GRouPcoin too, the people who had that runnign in EC2 instances to be used in such hardcoding in scripts seem to have shut down their GRouPcoin daemons...

Both DeVCoin and GRouPcoin could also do with having the DNS system set up, problem with both is I haven't yet any list of reliable nodes to actually plug into DNS for them...

-MarkM-
sr. member
Activity: 459
Merit: 250
I have two CLC connections too, and presumably luke-jr is no longer merged mining it with Eligius miners' hashing power as I have also got some confirmed blocks. Hmm or he is and I too am on a fork?

Maybe you are one of my two connections and I am one of your two? 142.177.234.* is one of the class C networks my ISP assigns me an IP on, and is the one I am on right now. Got a connection from that net?

-MarkM-


Nope, you're not...

ed@funkylinux4:~/apps$ netstat -n -p | grep coil
(Not all processes could be identified, non-owned process info
 will not be shown, you would have to be root to see it all.)
tcp        0      0 192.168.50.14:58509     99.162.89.???:8368      ESTABLISHED 1233/coiledcoind
tcp        0      0 192.168.50.14:36133     78.47.187.???:5330      ESTABLISHED 1233/coiledcoind
tcp        0      0 192.168.50.14:38177     78.129.222.???:8368     ESTABLISHED 1233/coiledcoind
tcp        0      0 192.168.50.14:51959     192.168.50.51:8368      ESTABLISHED 1233/coiledcoind
tcp        0      0 192.168.50.14:56598     173.246.103.???:6667     ESTABLISHED 1233/coiledcoind


Should we try to merge ourselves back together or is that even possible not knowing how long we've been seperated?
legendary
Activity: 1442
Merit: 1000
Sounds good Smiley P2pool is awesome Grin
hero member
Activity: 516
Merit: 643
So I'm still mining ok? (I upgraded to latest p2pool)

Yes. I'll push some changes soon that drop unupgraded peers instead of throwing up so many errors.
legendary
Activity: 1442
Merit: 1000
So I'm still mining ok? (I upgraded to latest p2pool)
hero member
Activity: 516
Merit: 643
I'm also getting the    "2012-02-26 02:31:41.256425 >     raise ValueError('old share an hour after switch time')
2012-02-26 02:31:41.256481 > exceptions.ValueError: old share an hour after switch time"    error message Sad

This is normal. Some other people haven't upgraded P2Pool and are messing with the rest..
legendary
Activity: 1442
Merit: 1000
I'm also getting the    "2012-02-26 02:31:41.256425 >     raise ValueError('old share an hour after switch time')
2012-02-26 02:31:41.256481 > exceptions.ValueError: old share an hour after switch time"    error message Sad
legendary
Activity: 2940
Merit: 1090
Yike the logs were over 32 megs, I emptied them in case sheer size of log was causing some of the problems I am seeing.

Git pull claims I am up to date with github.

The other day I saw litecoin has 124 or more connections, and killing it suddenly made my bandwidth stop lagging, so I figured its huge number of connections must be what was throttling my internet connection.

I set maxconnections=8, and it seems that is what has led to p2pool getting disconnected, since it gets 8 outgoing connections and subtracts that from total to see if it can also fit any incoming connections.

None of the other *coin daemons show any signs that the UPNP stuff works, plus in any case getinfo claims its not activated. So I wonder if my bandwidth was mere coincidence and actually my litecoind is not more reachable from the net on its p2p port than any other *coin daemon and maybe p2pool was somehoe those 116 or so incoming connections?

I am trying again now with maxconnections=9

I am also seeing old share over an hour old errors...

2012-02-26 02:31:41.256425 >     raise ValueError('old share an hour after switch time')
2012-02-26 02:31:41.256481 > exceptions.ValueError: old share an hour after switch time

All this is mining litecoins, the merged mining looks fine.

-MarkM-
hero member
Activity: 516
Merit: 643
Hmm heading off to check log, I just noticed I am now constantly getting "2012-02-26 01:51:47.863969 Bitcoin connection lost. Reason: Connection was closed cleanly." and/or not cleanly. Often its non-cleanly once followed by cleanly twice. This is with my instance that is mining litecoins and no merged ming so there aren't three daemons its talking too it must be makign and breacking connections to litecoind over and over very fast or something but neither the litecoind nor the minerd outputs show any inidcation they thing anythign is wrong.

Hmm where does p2pool put its log? Offhand I don't see one in the current directory of the tasks running it nor in its own directory nor in /var/log ...

Have you updated P2Pool recently? Bitcoin/Litecoin changed their protocols on Feb 20. The log is in data//log
legendary
Activity: 2940
Merit: 1090
Hmm heading off to check log, I just noticed I am now constantly getting "2012-02-26 01:51:47.863969 Bitcoin connection lost. Reason: Connection was closed cleanly." and/or not cleanly. Often its non-cleanly once followed by cleanly twice. This is with my instance that is mining litecoins and no merged ming so there aren't three daemons its talking too it must be makign and breacking connections to litecoind over and over very fast or something but neither the litecoind nor the minerd outputs show any inidcation they thing anythign is wrong.

Hmm where does p2pool put its log? Offhand I don't see one in the current directory of the tasks running it nor in its own directory nor in /var/log ...

(It is running in screen, which doesn't seem to have scrollback so all I can see of its output is one terminal windows worth at any given moment.)

-MarkM-
hero member
Activity: 516
Merit: 643
I am merged mining, BTC, NMC, IXC, I0C, DVC, GRP and CLC all at once.

I0coin has a much lower difficulty than IXCoin, yet I am seeing IXCoin blocks but no I0Coin blocks.

Does it actually submit the work to all the daemons it has enough difficulty for?

It should, though it hasn't been tested well. Can you check P2Pool's log for any mentions of "Merged block submittal"?
legendary
Activity: 2940
Merit: 1090
I have two CLC connections too, and presumably luke-jr is no longer merged mining it with Eligius miners' hashing power as I have also got some confirmed blocks. Hmm or he is and I too am on a fork?

Maybe you are one of my two connections and I am one of your two? 142.177.234.* is one of the class C networks my ISP assigns me an IP on, and is the one I am on right now. Got a connection from that net?

-MarkM-
Jump to: