Avalon users: bitcoind + eloipool configuration

[smracer]

root@ubuntu:/var/www/eloipool# python3 eloipool.py
2013-05-03 11:59:15,743 jsonrpc_getwork WARNING Error importing 'midstate' module; work will not provide midstates

>>> Traceback (most recent call last):
  File "eloipool.py", line 915, in <module>
    MM.start()
  File "/var/www/eloipool/merklemaker.py", line 697, in start
    self._prepare()
  File "/var/www/eloipool/merklemaker.py", line 120, in _prepare
    TS['access'] = URI2Access(TS['uri'])
  File "/var/www/eloipool/merklemaker.py", line 110, in URI2Access
    access = jsonrpc.ServiceProxy(uri)
AttributeError: 'module' object has no attribute 'ServiceProxy'


What am I doing wrong?

Run eloipool using runit.sh script..

#!/bin/sh

PYTHONPATH=/path/of/python-bitcoinrpc:/path/of/python-base58:/garz/repo/midstate \
     nohup ./eloipool.py 2>&1 >/dev/null &

I tried that.  Same error. 

[smracer]

Actually I get this error now that I fixed midstate

root@ubuntu:/var/www/eloipool# ./runit.sh

>>> Traceback (most recent call last):
  File "./eloipool.py", line 915, in <module>
    MM.start()
  File "/var/www/eloipool/merklemaker.py", line 697, in start
    self._prepare()
  File "/var/www/eloipool/merklemaker.py", line 120, in _prepare
    TS['access'] = URI2Access(TS['uri'])
  File "/var/www/eloipool/merklemaker.py", line 110, in URI2Access
    access = jsonrpc.ServiceProxy(uri)
AttributeError: 'module' object has no attribute 'ServiceProxy'

[jgarzik]

PYTHONPATH=/path/of/python-bitcoinrpc:/path/of/python-base58:/garz/repo/midstate \
     nohup ./eloipool.py 2>&1 >/dev/null &

Did you all change the "/path/of/" and "/garz/repo/" paths to point to the proper directory?

[smracer]

!/bin/sh

PYTHONPATH=/var/www/eloipool/python-bitcoinrpc:/var/www/eloipool/python-base58:/var/www/eloipool/midstate \
     ./eloipool.py

[intrd]

!/bin/sh

PYTHONPATH=/var/www/eloipool/python-bitcoinrpc:/var/www/eloipool/python-base58:/var/www/eloipool/midstate \
     ./eloipool.py

try to point jsonrpc too.. like this:

PYTHONPATH=$PYTHONPATH:$DIR/../python-bitcoinrpc/jsonrpc:$DIR/../python-bitcoinrpc/bitcoinrpc:$DIR/../python-base58:$DIR/../midstate \

in your case, replace $DIR variable to absolute path!

[intrd]

jgarzik:
this is normal? https://bitcointalk.org/index.php?topic=158105.msg2017114#msg2017114

[smracer]

!/bin/sh

PYTHONPATH=/var/www/eloipool/python-bitcoinrpc:/var/www/eloipool/python-base58:/var/www/eloipool/midstate \
     ./eloipool.py

try to point jsonrpc too.. like this:

PYTHONPATH=$PYTHONPATH:$DIR/../python-bitcoinrpc/jsonrpc:$DIR/../python-bitcoinrpc/bitcoinrpc:$DIR/../python-base58:$DIR/../midstate \

in your case, replace $DIR variable to absolute path!

Ok I pointed the jsonrpc as well and now i get this error.

root@ubuntu:/var/www/eloipool# ./runit.sh
Traceback (most recent call last):
  File "./eloipool.py", line 127, in <module>
    import jsonrpc_getwork
  File "/var/www/eloipool/jsonrpc_getwork.py", line 18, in <module>
    from jsonrpcserver import JSONRPCHandler
  File "/var/www/eloipool/jsonrpcserver.py", line 19, in <module>
    import json
  File "/var/www/eloipool/python-bitcoinrpc/jsonrpc/json.py", line 2, in <module>
    loads = _json.loads
AttributeError: 'module' object has no attribute 'loads'

[smracer]

Looking for a little guidance on eloipool setup.

I am fairly comfortable with ubuntu... though a bit new to python.

Anyway... I am running ubuntu 12.04 LTS server, and I think I have a stupid configuration error somewhere.  This is a pretty vanilla ubuntu 12.04 server with python3, libpython3.2, python3-dev, and build-essentials installed via apt.

I followed the top post and after all is said and done, I see in my eloipool.log the following message in a fairly tight loop:

You aren't doing anything wrong. I got the same thing. It seems to be a bug in the bitcoin-jsonrpc package, which is writing invalid basic-auth  headers out to bitcoind. It's setting the header to:

Code:

Basic b'....base64stuff...' 

rather than

Code:

  Basic ....base64stuff...

likely because it wasn't written for python3. I have no idea how this code would have ever worked, unless someone is running their bitcoind wide open without RPC passwords, or is using a different sort of authentication.

A quick fix is to decode the bytes into UTF-8. On line 72 of bitcoinrpc/authproxy.py change:

Code:

self.__auth_header = "Basic %s" % base64.b64encode(authpair)

to read:

Code:

self.__auth_header = "Basic %s" % base64.b64encode(authpair).decode()

Hope this helps

Solved my problem!  TY!!!

[optimator]

I have eloipool up and hashing away with an avalon unit. Thanks Jgarzik!! Very helpful guide.

I'm running into a bunch of errors relating to longpool - These errors are what has been described previously in this thread.

These errors appear non-fatal. Can anyone confirm this is correct?

[jspielberg]

I have been playing around with slush's pool software while waiting for B3 to go into production, and I like it as it has a handy little status webpage.

My preference though is eloipool.

Does eloipool have a method of checking the current running status of the pool?

Is the DB schema documented anywhere if I wanted to make my own scripts for status querying?

[Luke-Jr]

I have been playing around with slush's pool software while waiting for B3 to go into production, and I like it as it has a handy little status webpage.

My preference though is eloipool.

Does eloipool have a method of checking the current running status of the pool?

Is the DB schema documented anywhere if I wanted to make my own scripts for status querying?

Eloipool doesn't have a db schema... you have to configure that yourself.

[zero-asic]

I have been playing around with slush's pool software while waiting for B3 to go into production, and I like it as it has a handy little status webpage.

My preference though is eloipool.

Does eloipool have a method of checking the current running status of the pool?

Is the DB schema documented anywhere if I wanted to make my own scripts for status querying?

Eloipool doesn't have a db schema... you have to configure that yourself.

Correct me if I'm wrong here, but if you edit the ShareLogging section of the config.py as follows:

Code:

# Logging of shares:
ShareLogging = (
	#{
	#	'type': 'logfile',
	#	'filename': '../logs/share-logfile',
	#	'format': "{time} {Q(remoteHost)} {username} {YN(not(rejectReason))} {dash(YN(upstreamResult))} {dash(rejectReason)} {solution}\n",
	#},
	{
		'type': 'sql',
		'engine': 'mysql',
		'dbopts': {
			'host': 'localhost',
			'db': 'pooldb',
			'user': 'mysqluser',
			'password': 'mysqlpassword',
		},
		'statement': "insert into shares (rem_host, username, our_result, upstream_result, reason, solution, time) values ({Q(remoteHost)}, {username}, {YN(not(rejectReason))}, {YN(upstreamResult)}, {rejectReason}, unhex({solution}), NOW())",
	},
)

That should add DB entries to a table called shares.  It has fields: rem_host, username, our_result, upstream_result, reason, solution, time.

You could also do this with postgres I believe.  I'm not sure if this will make entries for all shares or just found blocks, I haven't tested it that much.

[Luke-Jr]

Try to use this:

Code:

pkill -f -USR1 eloipool.py

I'm guessing the problem is caused by running

Code:

... python3 eloipool.py ...

instead of

Code:

... ./eloipool.py ...

[Luke-Jr]

Now, if only you could tell me what is the significance and consequence of the multiple Eloipool log entries like this:

JSONRPCServer    INFO    Nobody to longpoll

No GBT/getwork longpoll connections were connected during a longpoll...

and

merkleMaker     INFO    Transaction-longpoll requested 102 seconds ago, and still not ready. Is your server fast enough to keep up with your configured WorkQueueSizeRegular maximum?

Your CPU isn't able to maintain the WorkQueueSize* configurations.
You can probably set all these to 1 since you don't use GBT/getwork with an Avalon.

[mrb]

You can probably set all these to 1 since you don't use GBT/getwork with an Avalon.

You mean Arvicco specifically did not configure it to use GBT/getwork, or cgminer on Avalon is by default is configured to not use it (ie. use stratum?)

[Luke-Jr]

You can probably set all these to 1 since you don't use GBT/getwork with an Avalon.

You mean Arvicco specifically did not configure it to use GBT/getwork, or cgminer on Avalon is by default is configured to not use it (ie. use stratum?)

I meant the former (that it was not in use, as evidenced by his first message), but cgminer's GBT is known to be broken, and AFAIK there aren't any cases where you'd actually want to use getwork on it (cgminer's getwork can't handle over a few Gh/s without rollntime, which Deepbit doesn't support; I don't know any other getwork-only pools)

[miner49dk]

Hey!
Thanks so much for a good walkthrough midstate was really a ***** to get complied on cents. Went ubuntu instead.
But
I have everything up and running, it seems to run perfectly, i do get a lot of "INFO Nobody to long poll" in the eloipool log.
Does this meen that there is something wrong? Am i actually Mining ?
The share log shows that the cgminer's i have running are working.
Hope you can help!

All the best
Ras

[Luke-Jr]

midstate is not needed for anything except DiabloMiner.

[miner49dk]

Hey !
Thanks, good to know.
Can you shed any ligt on the "nobody to longpoll" info in the logging ?
All the best
Ras

[tryexcept]

Hey !
Thanks, good to know.
Can you shed any ligt on the "nobody to longpoll" info in the logging ?
All the best
Ras

No GBT/getwork longpoll connections were connected during a longpoll...
If you just use avalons, since avalons use stratum, this is perfectly normal.

I have a question: What does it look like when you do find a block other than the address you configured will receive the block and fees ?
I didn't try test net although everything in the logs looks good, running on latest arch.