Backup? We don't need no steenking backup!   i'll ever have a backup server. i have seen to many days with all pools above 100% cdf. so i just don't care about "even out"-graphs |
|
|
|
I think I have read about the "DoS-like appearance" in some other thread before, can't remember where though.
But I don't think I have seen invalid shares being part of it, aggressive pulling/getwork I think it was when I read about it before.
The big question, is the massive amount of invalid shares only occurring on the ABCpool?
i saw two differences with abcpool to other pools: abcpool is lagging out very fast (only with bh) and my hashrate has gone crazy (but seems feasible if they are counting invalids for their hashrate calculation) i do think it is about bh's delag call; but i dont know how yet. it seems to be identic to a normal getwork. the delag calls are send every ten seconds; and thats really not much. |
|
|
|
To all other Bithopper users: I recommend to switch clients or to use another pool while this issue persists to avoid excess strain on our systems.
i told them in their forum. i think you could be a very good backup pool  looking forward to get things sorted |
|
|
|
the owner of abcpool just let us know not to use them atm: https://bitcointalk.org/index.php?topic=33586.msg457579#msg457579there are problems with counting invalids (you get your payment though), but for them bh behaves a little dos-sing. we don't know why yet; but we'll see. after that: it seems to be a really good pps pool as a backup |
|
|
|
i disabled you for now (did my login at work and forgot to take account details to my home...) tomorrow i'll look into bitHopper proxy code i'll see if i find any hint (i won't switch my client, but i really don't want to ddos you . as far as i know bitHopper it builds up the following connections to you: LP Call: (no data transmitted, connection will stay alive) getwork: {'method':'getwork', 'params':data, 'id':i} (i is a number just counting up, data is NULL or a submitted share) (approx 0.5/sec - LESS than with my three single miners connected) delag call:same as getwork (approx every ten seconds if your pool is marked as lagged) bithopper does mark your server as lagged if: the getwork returns NULL or the call did not get to your server properly. with all other pools it works flawlessly (if you allow me ddossing you for a few minutes) - tomorrow evening i'll look a little deeper. maybe i can tell you what makes you think of an invalid. btw: eligius does count stale and invalids also. with them i never get any invalids and very few rejects. so i still think its on your side. |
|
|
|
|
i think you have a bug with the invalid counter on your page.
my shares and rejects are counted correctly.
but my invalid rate goes higher and higher....(i use bithopper: but it gets higher even when it is connected to a different pool - bithopper saw you lagging - maybe its the way it is trying to "delag" you?)
this are my current stats:
Personal Stats
Current Hashrate: ~3200 MH/s
Lifetime Valid: 72 shares
Lifetime Stale: 1 shares
Lifetime Invalid: 619 shares
Lifetime Earned: 0.00189381 BTC
Balance: 0.00189381 BTC
Updated: 14:41:02 UTC
(its not a real problem as my shares and stales are counted correctly). but my hashrate is definatly NOT 3.2GH (damn....would be very nice - but its just around 1gh)
pool account name: flower
|
|
|
|
heya, any date when you'll pay us? and will you pay for your last 4min block - i did donate so i think i should get this money even if it was an invalid. btw. i really loved your pool. as a developer i was impressed how your pool did handle the load-changes from pool hoppers i hope you'll continue your work (maybe with smpps or something?) as i do love your design and your stability. if you want to open up again and need some help with a few nice graphs - just pm me |
|
|
|
You could however release a special version that does something like this and then provide a web service for everyone else to get the results of the consensus (as you don't need ~1000 samples).
perhaps can we use this: http://miner.k1024.de/#i dont have enough time to maintain that site properly. but (as i said way earlier) if some people would like to help me i could get them access to some of the openvz instances which do the stats grabbing. |
|
|
|
Ob es wohl Leute gibt, die einen MtGox-Code gegen Euro-Überweisung abkaufen würden? Hm.
Du kannst ihn bei bitcoin7.con einzahlen... mit referral: https://www.bitcoin7.com/?ref=6779b7 unterstützt euro handel und sepa. |
|
|
|
|
as a german user i just have to say:
paxxum is not an alternative. there fees are too high. PLEASE consider a bank account in europe. it does not cost very much and it could be done without ANY fees.
atm i'll stick to b7
|
|
|
|
|
no, you shouldn't see any miner idle warnings.
miner idle means your miner was not able to get more work.
would like to now if you see this error with stock poclbm too.
|
|
|
|
|
strange. i dont have any issues with them.
what do you see if you run a normal poclbm with them? i just added hopping code, i did not even touch the normal poclbm pool network code.
|
|
|
|
I'm getting lots of error's like this (on Ubuntu 10.10 64 bit):
/tmp/OCLgKNyWT.cl(197): error: bad argument type to opencl builtin function:
expected type "uint2", actual type "int"
W[27] = P1(27) + P3(27);
^
Error limit reached.
100 errors detected in the compilation of "/tmp/OCLgKNyWT.cl".
Compilation terminated.
i have answered here: https://bitcointalk.org/index.php?topic=34514.msg437154#msg437154 |
|
|
|
I'm getting lots of error's like this (on Ubuntu 10.10 64 bit):
/tmp/OCLgKNyWT.cl(197): error: bad argument type to opencl builtin function:
expected type "uint2", actual type "int"
W[27] = P1(27) + P3(27);
^
Error limit reached.
100 errors detected in the compilation of "/tmp/OCLgKNyWT.cl".
Compilation terminated.
hm, that does look like a poclbm error. could you try if this https://github.com/progranism/poclbm poclbm works for you? (i used it as a base; this error seems to be not related to my changes) |
|
|
|
Data point on biclockers:
With Cherry: .513% stales.
2.3Gh/s on one worker from 10 GPUs
Same account and worker since before I started hopping.
FWIW.
with my poclbm mod i am running at 0.28% and 0.33% stales (only twominers running). and still using the nicer stats grabbing and pool decisions from bitthopper Cool I'll have to check it out. Dies it work with ryouiki fork or just c00w's? I think it reinforces that bithopper is still identifiable by bitclockers when it's in between. i think it should work with ryo too.. never checked; but he forked after i integrated (my lame) first website - so i guess the json data could still be the same. just try |
|
|
|
I can't run your poclbm: C:\Python27\Lib\flower1024-poclbm-b77e028>"C:\Python27\python.exe" -O "C:\Python27\Lib\flower1024-poclbm-b77e028\poclbm.py" http://localhost:8337/data -d 1 -v -w64 -f10 --phatk2
07/08/2011 11:24:14, reading serverlist from bitHopper http://localhost:8337/data
07/08/2011 11:24:14, adding server user7516.1 @ arsbitcoin.com:8344
07/08/2011 11:24:14, adding server user7516.1 @ bitcoinmonkey.com:8332
07/08/2011 11:24:14, adding server user7516501 @ polmine.pl:8347
07/08/2011 11:24:14, adding server user7516.1 @ api2.bitcoin.cz:8332
07/08/2011 11:24:14, adding server user7516_1 @ digbtc.net:8332
07/08/2011 11:24:14, adding server user7516 @ mtred.com:8337
07/08/2011 11:24:14, adding server mSzTUg @ bitcoins.lc:8080
07/08/2011 11:24:14, adding server user7516 @ bitcoinpool.com:8334
07/08/2011 11:24:14, adding server user7516_1 @ pool.rfcpool.com:8332
Traceback (most recent call last):
File "C:\Python27\Lib\flower1024-poclbm-b77e028\poclbm.py", line 70, in <module>
miner = BitcoinMiner(devices[options.device], options, VERSION, HttpTransport.HttpTransport)
File "C:\Python27\Lib\flower1024-poclbm-b77e028\BitcoinMiner.py", line 45, in __init__
self.transport = transport(self, self.options.phatk2 or self.options.phatk2_1)
File "C:\Python27\Lib\flower1024-poclbm-b77e028\HttpTransport.py", line 20, in __init__
super(HttpTransport, self).__init__(miner)
File "C:\Python27\Lib\flower1024-poclbm-b77e028\Transport.py", line 53, in __init__
say_line("adding server " + js["servers"][srv]["user"] + " @ " + js["servers"][srv]["mine_address"])
KeyError: 'mine_address' hey, it looks like your bithopper configuration lacks the mine_address for rfcpool. just look in bithoppers pools.py if all pools have a valid mine_address |
|
|
|
it's working now.. still get that message and some miner idle messages but it seems to be running well now. yeah might have spoken too soon... bithopper says i am running at 0mh and isnt updating any shares but you screen seems to suggest shares are being done bmunion 07/08/2011 02:21:42, warning: job finished, miner is idle
bmunion [312.977 MH/s (~198 MH/s)] [Rej: 1/18 (5.56%)] but as you can see i get the idle errors. as is said: shares can't be reported to bithopper as all is submitted directly. |
|
|
|
we can pester sukrim he made one for bithopper.
i'm nearly there i think.. downloading visual studios express 2008.
the biggest problem is pyopencl on win7 x64 but i can't support you on that as i never got pyopencl it working on my win7 x64 machine (didn't try very hard though) |
|
|
|
well first it told me it was missing numby.. i googled and installed that.
they it said it was missing pyopencl and being on windows, it told me i wasnt going to finish this tonight
its just a classic poclbm with a few addons. i guess i should make a windows exe (this would include all the needed dependencies) - but i never done that before. on linux it should run flawlessly |
|
|
|
|
Show Posts
