New stratum/getwork proxy with Web-based GUI and pools management

[girino]

The pool connection problem you describe sould not be caused by the connection limit. If you can retrieve the logs when this problem comes back, please send me the file.


When the limit number of connections is reached, the connection is rejected with the following error in the logs:

Too many connections on pool xxxx for the connection xxxxxx. Sending error and close the connection.


On the client side, the following error should be displayed in the mining software:

Too many connection on the pool.

Hum, ok.

What about this log here:

Code:

2015-01-08 06:06:10,673 ERROR [Pool-Nicehash X11-Thread]: [Pool]    Disconnect of pool Pool [name=Nicehash X11, host=stratum.nicehash.com:3336, username=xxx, password=x, readySince=Thu Jan 08 06:01:15 EST 2015, isReady=true, isEnabled=true, isStable=true, priority=0, weight=9].
java.io.EOFException: Connection closed.
        at strat.mining.stratum.proxy.network.StratumConnection$1.run(StratumConnection.java:164)
2015-01-08 06:06:10,681 WARN  [Pool-Nicehash X11-Thread]: [ProxyManager]    Pool Nicehash X11 is DOWN. Moving connections to another one.
2015-01-08 06:06:10,682 INFO  [Pool-Nicehash X11-Thread]: [MonoCurrentPoolStrategyManager]    Switching worker connections from pool Nicehash X11 to pool Westhash X11.
2015-01-08 06:06:10,684 INFO  [Pool-Nicehash X11-Thread]: [MonoCurrentPoolStrategyManager]    Close connection /112.111.226.142:13474 since the on-the-fly extranonce change is not supported.

Not that it affects much, clients are reconnected just fine, but if that happens a lot it can lower my hashrate.

[1714953820]

1714953820

[1714953820]

1714953820

[Pfool]

Could you post the 50 lines above these ones. I think the cause is in these lines...

[girino]

Could you post the 50 lines above these ones. I think the cause is in these lines...

all of the 100 or so previous lines are of the sort:

Code:

2015-01-08 06:02:01,169 INFO  [Pool-Nicehash X11-Thread]: [ProxyManager]    New block detected on pool Nicehash X11.

nothing unusual at all. I just disabled "enableExtranonceSubscribe". Let's hope it helps.

[Anon39]

it worked well for a few hours and suddenly it stops showings a lot of errors like this:

Code:

2015-01-08 x:x:x,122 INFO     [TimerExecutorThread-8]:strat.mining.stratum.proxy.manager.strategy.MonoCurrentPoolStrategyManager - Close connection /xx.xx.xx.xxx:11952 since the on-the-fly extranonce change is not supported.

and miningrigrentals stop the miner saying "sick pool".
i tryed "enableExtranonceSubscribe" : false but i get the same error.

[Amph]

it worked well for a few hours and suddenly it stops showings a lot of errors like this:

Code:

2015-01-08 x:x:x,122 INFO     [TimerExecutorThread-8]:strat.mining.stratum.proxy.manager.strategy.MonoCurrentPoolStrategyManager - Close connection /xx.xx.xx.xxx:11952 since the on-the-fly extranonce change is not supported.

and miningrigrentals stop the miner saying "sick pool".
i tryed "enableExtranonceSubscribe" : false but i get the same error.

maybe not enough ram?

can this proxy be optimized to consume less ram, or it's the nature of jave to be so ram hungry?

[Anon39]

it worked well for a few hours and suddenly it stops showings a lot of errors like this:

Code:

2015-01-08 x:x:x,122 INFO     [TimerExecutorThread-8]:strat.mining.stratum.proxy.manager.strategy.MonoCurrentPoolStrategyManager - Close connection /xx.xx.xx.xxx:11952 since the on-the-fly extranonce change is not supported.

and miningrigrentals stop the miner saying "sick pool".
i tryed "enableExtranonceSubscribe" : false but i get the same error.

maybe not enough ram?

can this proxy be optimized to consume less ram, or it's the nature of jave to be so ram hungry?

it is working on a 4gb ubuntu desktop... 2,5gb free ram

[Pfool]

I will need your complete log file to help you.

How much RAM the proxy take on your computer ? You can try to lower the RAM consumption by adding the -XmxYYYM before the -jar option on the command line where YYY is the max memory size in MB for the java heap space (the java process will still consume a bit more than this specified size).

You can try with a 64MB size.

[bleeding2323]

Yea! Success!. We'll see how it goes. Ok here is how it works:


I have 2 batch files called proxy_black.bat (for the 25 mHs)
and                              proxy_blades.bat (for the 10 Mhs total)

proxy_black.bat command:
java -jar stratum-proxy.jar -f stratum-proxy-nicehash.conf

proxy_blades.bat command:
java -jar stratum-proxy.jar -f stratum-proxy-nicehash2.conf

First instance config file:

stratum-proxy-nicehash.conf:

{
  "isScrypt" : true,
  "disableStratum" : true,
  "pools" :
  [
  {
    "name" : "Nicehash Scrypt",
    "host" : "stratum.nicehash.com:3333",
    "user" : "<your bitcoin address.black", <----you can rename "black" to what you want just make sure there is a "." seperating the bitcoin address
    "password" : "x",
    "enableExtranonceSubscribe" : true
  },
  {
    "name" : "Westhash Scrypt",
    "host" : "stratum.westhash.com:3333",
    "user" : "your bitcoin address.black",
    "password" : "x",
    "enableExtranonceSubscribe" : true
  }
  ]
}


Second config file:

stratum-proxy-nicehash2.conf:

{
  "getworkListenPort" : 8333,               <----different port number to connect your miners ie http://proxyip:8333
  "databaseDirectory": "database1",      <----different directory for database must be different for each instance
  "isScrypt" : true,
  "disableStratum" : true,
  "pools" :
  [
  {
    "name" : "Nicehash Scrypt2",
    "host" : "stratum.nicehash.com:3333",
    "user" : "1your bitcoin address.blades",   <-------different worker name for second instance
    "password" : "x",
    "enableExtranonceSubscribe" : true
  },
  {
    "name" : "Westhash Scrypt2",
    "host" : "stratum.westhash.com:3333",
    "user" : "your bitcoin address.blades",
    "password" : "x",
    "enableExtranonceSubscribe" : true
  }
  ]

There you go you can just keep creating instance as long as you give each instance there own directory for the database file. Just create database1, database2, etc and copy the dbpools and dbusers from the original database directory.

how would you set this up for sha256, and scrypt, on one instace isscrypt false?, and on the next iscrypt true, and of course different ports

[Pfool]

how would you set this up for sha256, and scrypt, on one instace isscrypt false?, and on the next iscrypt true, and of course different ports

To run several instances of the proxy on a same host, you have indeed to change (or disable) the 3 ports for stratum, getwork and API, but you also have to set a different databaseDirectory for each instance through the "databaseDirectory" directory parameter.

[Amph]

I will need your complete log file to help you.

How much RAM the proxy take on your computer ? You can try to lower the RAM consumption by adding the -XmxYYYM before the -jar option on the command line where YYY is the max memory size in MB for the java heap space (the java process will still consume a bit more than this specified size).

You can try with a 64MB size.

so example

-Xmx64M for 64mb?

are 64mb enough to run it?

[Pfool]

so example

-Xmx64M for 64mb?

Yes

so example

are 64mb enough to run it?

It depends of the number of workers connected on the proxy. I have had the proxy run with only 16mb with 4 workers. But you should experiment several values. It also depends if you run the proxy on a 32 bits or 64 bits JVM (64 bits will need more memory).

If the memory is too low, you will surely see some MemoryError in the logs or see a high CPU usage of the proxy.

[Amph]

so example

-Xmx64M for 64mb?

Yes

so example

are 64mb enough to run it?

It depends of the number of workers connected on the proxy. I have had the proxy run with only 16mb with 4 workers. But you should experiment several values. It also depends if you run the proxy on a 32 bits or 64 bits JVM (64 bits will need more memory).

If the memory is too low, you will surely see some MemoryError in the logs or see a high CPU usage of the proxy.

i'm running it with 4 workers yes, and it give me plenty of error sometimes, refuse connection/closed connection, after that it stuck there and i must restart the whole vps

[Pfool]

so example

-Xmx64M for 64mb?

Yes

so example

are 64mb enough to run it?

It depends of the number of workers connected on the proxy. I have had the proxy run with only 16mb with 4 workers. But you should experiment several values. It also depends if you run the proxy on a 32 bits or 64 bits JVM (64 bits will need more memory).

If the memory is too low, you will surely see some MemoryError in the logs or see a high CPU usage of the proxy.

i'm running it with 4 workers yes, and it give me plenty of error sometimes, refuse connection/closed connection, after that it stuck there and i must restart the whole vps

You have to restart the VPS when the proxy craches ? 
I rarelly saw a simple process craching a whole system since Windows 98... (and never on Linux)

[bleeding2323]

my proxy is up, and working great.
I can not connect my miners to it, updated cgminer on one of my antminer s3 to 4.6.1,  this one still will not connect.

do I need to assign my proxy a port, then connect my miners to that port?
lol I got it, just was not reading  

its connected to the westhash pool, just not showing any hash..
now its showing all rejects....cgminer is updated, and bfgminer is updated as needed, I'm missing something

[bleeding2323]

how would you set this up for sha256, and scrypt, on one instace isscrypt false?, and on the next iscrypt true, and of course different ports

To run several instances of the proxy on a same host, you have indeed to change (or disable) the 3 ports for stratum, getwork and API, but you also have to set a different databaseDirectory for each instance through the "databaseDirectory" directory parameter.

Im not sure how to do that, is there a way you could show me what I need to do...this is still all new to me, I like it, but still dumb to it.
I am using asics, so I know the conf will be diferent,

[Amph]

so example

-Xmx64M for 64mb?

Yes

so example

are 64mb enough to run it?

It depends of the number of workers connected on the proxy. I have had the proxy run with only 16mb with 4 workers. But you should experiment several values. It also depends if you run the proxy on a 32 bits or 64 bits JVM (64 bits will need more memory).

If the memory is too low, you will surely see some MemoryError in the logs or see a high CPU usage of the proxy.

i'm running it with 4 workers yes, and it give me plenty of error sometimes, refuse connection/closed connection, after that it stuck there and i must restart the whole vps

You have to restart the VPS when the proxy craches ? 
I rarelly saw a simple process craching a whole system since Windows 98... (and never on Linux)

yeah it stuck there, ctrl+c doesn't work

[girino]

Hi,

i've been using this proxy for quite sometime and it is great. I just need to know a little more about this:

-Up to 256 workers at the same time (stratum and getwork)

1- is this the number of workers, or the number of connections? Can I have a single worker with 300 connections, for example?
2- is this a restriction of the architecture or can this be changed? (I don't mind having to compile the code to increase it, just need to know where to change it).

I have being looking into to the code and have not seen this restriction coded anywhere, that's why i asked question 2.

Thanks in advance,
Girino.

This is a limit on the number of connections.

You can try to increase this limit to 65536 connections by modifying a constant in the source code. It should work but I never tested it. The drawback is that you may have problems with Nicehash. It should be rare but it may happen depending on the pool of the job you will work on (But you have no control on that). The symptom will be a disconnection of the pool with the message "The extranonce2Size for the pool nameOfThePool is too low. Size: X, mininum needed Y.". But on the next reconnection of the pool (if you are lucky), the problem will vanish since Nicehash will affect you on a different job.

Change the value of the constant DEFAULT_EXTRANONCE1_TAIL_SIZE in the Constants.java file from 1 to 2.

In all cases, I am interested by the result of your experiment.
Good luck.

Got this error today for the first time (2 days running) and apparently it didn't really affect my hashrate:

Code:

The pool extranonce2 size is too low (2). Minimum is 3

so i ' m probably not rolling back the change unless it occurs more frequently. Its working great for the moment.

[chrysophylax]

how would you set this up for sha256, and scrypt, on one instace isscrypt false?, and on the next iscrypt true, and of course different ports

To run several instances of the proxy on a same host, you have indeed to change (or disable) the 3 ports for stratum, getwork and API, but you also have to set a different databaseDirectory for each instance through the "databaseDirectory" directory parameter.

exactly what i have done ...

one server and 6 proxies on the same server ...

different ports - different pools - different log directories ...

works like a charm ... just awesome ...

#crysx

[chrysophylax]

how would you set this up for sha256, and scrypt, on one instace isscrypt false?, and on the next iscrypt true, and of course different ports

To run several instances of the proxy on a same host, you have indeed to change (or disable) the 3 ports for stratum, getwork and API, but you also have to set a different databaseDirectory for each instance through the "databaseDirectory" directory parameter.

Im not sure how to do that, is there a way you could show me what I need to do...this is still all new to me, I like it, but still dumb to it.
I am using asics, so I know the conf will be diferent,

in the config file - change the ports ( i changed all of them to the next consecutive one - eg proxy 1 - 3333 / 8332 / 8888 ... proxy 2 - 3334 / 8333 / 8889 ) and also set the log file ( settign right at the beginning of the conf file ) to a different directory for EACH of the proxies ... that way - you dont overlap / overwrite any of the log files ...

im using 0.5.1 release ( not compiled ) on a scientific linux 6.5 server ( rhel linux distro ) with 16GB high performance ram and 24TB RAID storage ( its also a data storage server ) and it run very well without a flaw ... each of the 6 proxies are pointing / connecting to a different pool and the miners ( a mix of ccminer and sgminer ) are mining / pointed at different proxies on the same host / server ...

eg miner 1 - ./sgminer -k x11 stratum+tcp://192.168.0.205:3333 etc ( proxy 1 - granitecoin ) ... miner 2 - ./sgminer -k neoscrypt stratum+tcp://192.168.0.205:3334 etc ( proxy 2 - yaamp neoscrypt ) ... miner 3 - ./ccminer -a lyre2re stratum+tcp://192.168.0.205:3335 etc ( proxy 3 - westhash lyre2re ) ... etc ...

all these miners are using stratum and none using getwork ( port 8332 - proxy 1 ... port 8333 - proxy 2 ... port 8334 - proxy 3 ... etc ) ...

very very very stable ...

i actually have 3 of the proxies pointed to different mining pools for granitecoin ... https://coinking.io ( which is setup for grn / x11 ) - http://aikapool.com/grn/ - http://www.zpool.co/pool/grn/x11 ...

so in sgminer ( unfortunately not for ccminer - yet ( i hope ) ) i just use --failover-only and assign the next proxy thats on the list ...

the other miners i use for multipool yaamp and westhash ... and one proxy that i change often ( which has multiple conf files ) depending on what i want to mine for the rest of the workers in the farm ...

hope this helps ...

#crysx

[girino]

Tried to recompile the latest version from github and got the following error in the "log" pane from the web interface (click to enlarge):