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

[mdude77]

I upgraded to the stratum enabled p2pool & cgminer but had to turn stratum off due to massive rise in stale/rejects, even after lowering my intensity on both Linux & windoze rigs - anyone else having these problems? 

did you upgrade to latest cgminer?

Yup, latest of both.....

well then just use getwork, unless u got more than 50 GH/s it dosnt need stratum

Done. Much better results. I was of the understanding that stratum was going to be an improvement ie: less stales/DOA?

it should yes, altough if there are problems (since its new and therefore probably bit buggy) u can get worse performance. im running cminer too and i got 0.36% stale, all my shares have been accepted (due to alot of connections) so far with stratum enabled.

I had to disable stratum here as well, higher rejects with stratum than with LP.

M

Interesting indeed. I'm running cgminer/p2pool with Stratum and am not having the issues you and others are mentioning. My reject rate is 2.3% and is the same as it was when I was running getwork. I still show Zero stales on Stratum as I did on getwork as well.

p2pool and bitcoind are running on a local server on my LAN. My rigs communicate with them on that same LAN/subnet.
 
I do not use the un-released git version of bitcoind.

If you are using an un-released git version of bitcoind or are using a non local p2pool server that could be your problem.

I'm not using an unreleased version of bitcoind.  I tried it for a while, and it didn't help any.  I've come to the conclusion I don't have the bandwidth to run p2pool the "proper" way without getting excessive stales.  I've turned off port forwarding for everything - namecoind, bitcoind, and p2pool.  My stales are consistently low now, unless someone else starts hogging bandwidth on my network.

Tomorrow I think I'll try stratum again and see how things go.

M

[1713888988]

1713888988

[1713888988]

1713888988

[K1773R]

I upgraded to the stratum enabled p2pool & cgminer but had to turn stratum off due to massive rise in stale/rejects, even after lowering my intensity on both Linux & windoze rigs - anyone else having these problems? 

did you upgrade to latest cgminer?

Yup, latest of both.....

well then just use getwork, unless u got more than 50 GH/s it dosnt need stratum

Done. Much better results. I was of the understanding that stratum was going to be an improvement ie: less stales/DOA?

it should yes, altough if there are problems (since its new and therefore probably bit buggy) u can get worse performance. im running cminer too and i got 0.36% stale, all my shares have been accepted (due to alot of connections) so far with stratum enabled.

I had to disable stratum here as well, higher rejects with stratum than with LP.

M

Interesting indeed. I'm running cgminer/p2pool with Stratum and am not having the issues you and others are mentioning. My reject rate is 2.3% and is the same as it was when I was running getwork. I still show Zero stales on Stratum as I did on getwork as well.

p2pool and bitcoind are running on a local server on my LAN. My rigs communicate with them on that same LAN/subnet.
 
I do not use the un-released git version of bitcoind.

If you are using an un-released git version of bitcoind or are using a non local p2pool server that could be your problem.

I'm not using an unreleased version of bitcoind.  I tried it for a while, and it didn't help any.  I've come to the conclusion I don't have the bandwidth to run p2pool the "proper" way without getting excessive stales.  I've turned off port forwarding for everything - namecoind, bitcoind, and p2pool.  My stales are consistently low now, unless someone else starts hogging bandwidth on my network.

Tomorrow I think I'll try stratum again and see how things go.

M

see cgminer thread for my answer

[mdude77]

Been running with stratum now for close to 12 hours.  I've turned off port forwarding completely.  So far so good..

27051 accepted from cgminer, 544 rejected or 2.01% rejects. 

More importantly, though, my payout is where it's supposed to be.  Last time I lost 30% of my payout running stratum overnight.  It was with the pre-release v10 of p2pool, however, but I also had unlimited incoming connections.  Not sure which "fixed" it, but all seems to be well now.

M

[rav3n_pl]

grep SHARE p2pool/data/bitcoin/log
2012-12-31 07:26:37.637027 GOT SHARE! toy.gpu 005461de prev 7f752c8e age 1.00s
2012-12-31 09:41:22.081034 GOT SHARE! toy.gpu 5bb6c97b prev 28ae9b73 age 10.18s
2012-12-31 11:42:48.881662 GOT SHARE! toy.gpu 498af61a prev a690bd70 age 1.22s
2012-12-31 15:27:47.966547 GOT SHARE! 1FooBarwPJBCHENtafHz7UCfSK15FTh8wm 7c447bf4 prev 33c2552f age 0.78s
2012-12-31 16:17:29.271429 GOT SHARE! toy.gpu eaf9f045 prev 46eef6d5 age 1.84s

Shares: 5 total (5 orphaned, 0 dead) Efficiency: 0.000%

Mean GBT is 0.2s

WHY all orphaned???

something wrong with ur bitcoind

Recompiled to latest git and restarted, will see.
Nothing wrong i see in bitcoin log anyway.

Got  2 another shares and both orphaned.... I never had have SO BAD LUCK!

[zvs]

2013-01-01 08:32:06.699172 New work for worker! Difficulty: 1.565199 Share difficulty: 584.008722 Total block value: 25.481000 BTC including 420 transactions
2013-01-01 08:32:07.245097 GOT SHARE!  3cc1ec10 prev 43e63923 age 0.92s
2013-01-01 08:32:07.276257 > Share check failed:
2013-01-01 08:32:07.276323 > Traceback (most recent call last):
2013-01-01 08:32:07.276350 >   File "/home/jesus/p2pool/p2pool/bitcoin/worker_interface.py", line 70, in _getwork
2013-01-01 08:32:07.276384 >     defer.returnValue(self.merkle_root_to_handler[header['merkle_root']](header, request.getUser() if request.getUser() is not None else '', 0))
2013-01-01 08:32:07.276458 >   File "/home/jesus/p2pool/p2pool/work.py", line 345, in got_response
2013-01-01 08:32:07.276493 >     self.node.set_best_share()
2013-01-01 08:32:07.276536 >   File "/home/jesus/p2pool/p2pool/node.py", line 288, in set_best_share
2013-01-01 08:32:07.276569 >     best, desired, decorated_heads = self.tracker.think(self.get_height_rel_highest, self.bitcoind_work.value['previous_block'], self.bitcoind_work.value['bits'], self.known_txs_var.value)
2013-01-01 08:32:07.276610 >   File "/home/jesus/p2pool/p2pool/data.py", line 445, in think
2013-01-01 08:32:07.276641 >     if self.attempt_verify(share):
2013-01-01 08:32:07.276664 > --- <exception caught here> ---
2013-01-01 08:32:07.276690 >   File "/home/jesus/p2pool/p2pool/data.py", line 424, in attempt_verify
2013-01-01 08:32:07.276713 >     share.check(self)
2013-01-01 08:32:07.276746 >   File "/home/jesus/p2pool/p2pool/data.py", line 314, in check
2013-01-01 08:32:07.276778 >     raise ValueError('''gentx doesn't match hash_link''')
2013-01-01 08:32:07.276809 > exceptions.ValueError: gentx doesn't match hash_link


what exactly does that signify?


it would have kept going forever if i had not caught it

re:

log.old.2:2013-01-01 08:19:28.273919 > Share check failed:
log.old.2:2013-01-01 08:23:18.504064 > Share check failed:

would it be from using the latest git bitcoind?

    "version" : 79900,
    "protocolversion" : 60002,
    "walletversion" : 60000,
    "balance" : 0.00000000,
    "blocks" : 214663,
    "connections" : 959,
    "proxy" : "",
    "difficulty" : 2979636.61693807,
    "testnet" : false,
    "keypoololdest" : 1357027515,
    "keypoolsize" : 1,
    "paytxfee" : 0.00000000,
    "errors" : "This is a pre-release test build - use at your own risk - do not use for mining or merchant applications"


i suppose it does give a warning at the end

    "blocks" : 214663,
    "currentblocksize" : 274839,
    "currentblocktx" : 465,
    "difficulty" : 2979636.61693807,
    "errors" : "This is a pre-release test build - use at your own risk - do not use for mining or merchant applications",
    "generate" : false,
    "genproclimit" : -1,
    "hashespersec" : 0,
    "pooledtx" : 986,
    "testnet" : false



anyway, i set it back to 30k size, as the only time ive had that problem (i guess that's what the other thing with the 25 orphans was as well) is when im not limiting block size

[K1773R]

happens for me now too, how do i disable stratum?

[mdude77]

happens for me now too, how do i disable stratum?

Code:

--fix-protocol      Do not redirect to a different getwork protocol (eg. stratum)

M

[AV]

K1773R
cgminer
--fix-protocol      Do not redirect to a different getwork protocol (eg. stratum)


Ups...

[K1773R]

yea, was just about to EDIT my post since i found it... shame on me!
its funny, yesterday i had no problem at all...

[rav3n_pl]

Posted issue on git, looks like thats WHY all my shares being orphaned...
https://github.com/forrestv/p2pool/issues/73

[mdude77]

Posted issue on git, looks like thats WHY all my shares being orphaned...
https://github.com/forrestv/p2pool/issues/73

Bad build of bitcoind?

M

[rav3n_pl]

Posted issue on git, looks like thats WHY all my shares being orphaned...
https://github.com/forrestv/p2pool/issues/73

Bad build of bitcoind?

M

Trouble started after update to v10. I`m using git build of bitcoind, not have any issues earlier.

[lenny_]

I've been working with rav3n_pl to find out why my local Stratum mining proxy cannot connect to p2pool anymore, after upgrading to p2pool v10.
Finally we reported an issue on Stratum mining proxy: https://github.com/slush0/stratum-mining-proxy/issues/6
That's what slush respond:

slush0 commented a day ago

It is server side problem; server should send mining.notify message directly after mining.subscribe and not wait for mining.authorize.

Please ask server operator to fix it on his side.

Help needed

[rav3n_pl]

Lenny, it is fixed already
https://github.com/forrestv/p2pool/commit/2dffea00ee99f2c624a8504afee37532ba3daa5d

[btctalk]

I was about to reply with the same problem that saw the updated version. I just updated hope that it solved the issue, I had lots of orphaned shares on v10 so far...

btw is there any way to prevent the statictic's reset when restarting p2pool?

[lenny_]

Lenny, it is fixed already
https://github.com/forrestv/p2pool/commit/2dffea00ee99f2c624a8504afee37532ba3daa5d

Thanks for info I wasn't there for a while, you know, New Year's Eve

[rav3n_pl]

I was about to reply with the same problem that saw the updated version. I just updated hope that it solved the issue, I had lots of orphaned shares on v10 so far...

btw is there any way to prevent the statictic's reset when restarting p2pool?

Nope. Number of shares is counted from 0.
But rest of stats is in the graphs, they are not lost.

[forrestv]

P2Pool release 11.0 tag: 11.0 hash: bf3056c76330dcc776ad09650c567619e94dea3b

Windows binary: http://u.forre.st/u/fepwxaia/p2pool_win32_11.0.zip
Source zipball: https://github.com/forrestv/p2pool/zipball/11.0
Source tarball: https://github.com/forrestv/p2pool/tarball/11.0

Changes:
* Fixed shares being orphaned with error message, caused by a bug the subsidy fix in the previous release
* Work is shared between stratum connections, resulting in greatly decreased latency when many stratum clients are connected
* Fixed compatibility with stratum-mining-proxy

[PatMan]

Ah! Great work - and so quick! Upgraded, let's see how goes it. Many thanks.

[lenny_]

Stratum-mining-proxy still not working at all. At least, BTC Miner itself work now with default protocol GetWork without problems.
Here some logs:

1. Connected Ztex BTCMiner with BTC address as user to Stratum-mining-proxy
Stratum-mining-proxy error output:

Code:

2013-01-02 15:21:33,368 INFO proxy client_service.handle_event # Setting new difficulty: 0.999984741211
2013-01-02 15:21:33,373 INFO proxy client_service.handle_event # New job 218616616958190723782478161383154985014 for prevhash ccca5c12, clean_jobs=True
2013-01-02 15:21:41,314 WARNING proxy getwork_listener.render_POST # Worker '1JsdviS8oYHVtLxEzFYGHUTZk88EQtN1Bd' subscribed for LP
2013-01-02 15:21:43,670 INFO proxy client_service.handle_event # Setting new difficulty: 0.999984741211
2013-01-02 15:21:43,676 INFO proxy client_service.handle_event # New job 109043941172271690938874337452704092682 for prevhash ccca5c12, clean_jobs=True
2013-01-02 15:21:43,677 INFO proxy getwork_listener._on_lp_broadcast # LP broadcast for worker '1JsdviS8oYHVtLxEzFYGHUTZk88EQtN1Bd'
2013-01-02 15:21:43,678 INFO proxy jobs.extranonce2_padding # Extranonce size mismatch. Please report this error to pool operator!
2013-01-02 15:21:43,698 WARNING proxy getwork_listener.render_POST # Worker '1JsdviS8oYHVtLxEzFYGHUTZk88EQtN1Bd' subscribed for LP

Ztex BTCMiner output:

Code:

2013-01-02T15:22:23: 001-0: ztex_ufm1_15y1-04A39E424E-3: Error: jsonParse: Parameter `data' not found: Disabling URL http://localhost:4555 for 60s

2. Connected Ztex BTCMiner with empty user field to Stratum-mining-proxy
Startum-mining-proxy output:

Code:

2013-01-02 15:23:37,214 WARNING proxy getwork_listener.render_POST # Authorization required

Ztex BTCMiner output:

Code:

001-0: ztex_ufm1_15y1-04A3465EA1-1: Error: Server returned HTTP response code: 401 for URL: http://localhost:4555: Disabling URL http://localhost:4555 for 60s