pushpool - open source pool software

[turlando]

I found this log:

Code:

% tail /tmp/request.log
[2011-06-02 13:48:43.821358] ::ffff:192.168.1.210 - "/"
[2011-06-02 13:49:13.826827] ::ffff:192.168.1.210 - "/"
[2011-06-02 13:49:43.833494] ::ffff:192.168.1.210 - "/"
[2011-06-02 13:50:13.838898] ::ffff:192.168.1.210 - "/"
[2011-06-02 13:50:43.844413] ::ffff:192.168.1.210 - "/"
[2011-06-02 13:51:13.851388] ::ffff:192.168.1.210 - "/"
[2011-06-02 13:51:43.856926] ::ffff:192.168.1.210 - "/"
[2011-06-02 14:07:14.385247] ::ffff:192.168.1.210 - "/"
[2011-06-02 14:07:31.446938] ::ffff:192.168.1.210 - "/"
[2011-06-02 15:16:10.756269] ::ffff:192.168.1.210 - "/"

memcached is running properly:

Code:

% rc.d restart memcached
:: Stopping memcached                                                                                                               [DONE]
:: Starting memcached                                                                                                               [DONE]

I also added to /etc/hosts.allow:

Code:

pushpoold: ALL

[1714016868]

1714016868

[1714016868]

1714016868

[1714016868]

1714016868

[1714016868]

1714016868

[Jine]

I'm personally having scalability issues (i think?) with Deb+pushpoold.

We're serving ~200 workers over at Bitcoins.lc, and I'm getting errors like this all the time.

[2011-06-03 03:55:36] thread 2: 97416096 hashes, 1618.00 khash/sec
[2011-06-03 03:55:36] JSON-RPC call failed: {
   "code": -2,
   "message": "upstream RPC error"
}
[2011-06-03 03:55:36] json_rpc_call failed, retry after 30 seconds
[2011-06-03 03:55:37] thread 3: 97416096 hashes, 1624.76 khash/sec

Why is "bitcoind" refusing to give out work sometimes?
Hosted on a 8 CPU VPS with lots of ram.

Any suggestions?
Bounty for solving this problem is a couple of BTC, easily.

EDIT: Also:

Jun  3 03:57:01 bitcoins pushpoold[9380]: HTTP request failed: couldn't connect to host
Jun  3 03:57:01 bitcoins pushpoold[9380]: HTTP request failed: couldn't connect to host
Jun  3 03:57:01 bitcoins pushpoold[9380]: HTTP request failed: couldn't connect to host

[Jine]

I'm also seeing shitload of TIME_WAIT connections between the upstream and pushpoold:

tcp        0      0 127.0.0.1:8332          127.0.0.1:14446         TIME_WAIT   -
tcp        0      0 127.0.0.1:8332          127.0.0.1:14412         TIME_WAIT   -
tcp        0      0 127.0.0.1:8332          127.0.0.1:14276         TIME_WAIT   -
tcp        0      0 127.0.0.1:8332          127.0.0.1:14233         TIME_WAIT   -
tcp        0      0 127.0.0.1:8332          127.0.0.1:15052         TIME_WAIT   -
tcp        0      0 127.0.0.1:8332          127.0.0.1:14314         TIME_WAIT   -
tcp        0      0 127.0.0.1:8332          127.0.0.1:14667         TIME_WAIT   -
tcp        0      0 127.0.0.1:8332          127.0.0.1:14726         TIME_WAIT   -
tcp        0      0 127.0.0.1:8332          127.0.0.1:14822         TIME_WAIT   -
tcp        0      0 127.0.0.1:8332          127.0.0.1:14548         TIME_WAIT   -
tcp        0      0 127.0.0.1:8332          127.0.0.1:14499         TIME_WAIT   -
tcp        0      0 127.0.0.1:8332          127.0.0.1:14335         TIME_WAIT   -
tcp        0      0 127.0.0.1:8332          127.0.0.1:14677         TIME_WAIT   -
tcp        0      0 127.0.0.1:8332          127.0.0.1:13976         TIME_WAIT   -
tcp        0      0 127.0.0.1:8332          127.0.0.1:14348         TIME_WAIT   -
tcp        0      0 127.0.0.1:8332          127.0.0.1:14320         TIME_WAIT   -
tcp        0      0 127.0.0.1:8332          127.0.0.1:14949         TIME_WAIT   -
tcp        0      0 127.0.0.1:8332          127.0.0.1:13990         TIME_WAIT   -
tcp        0      0 127.0.0.1:8332          127.0.0.1:14439         TIME_WAIT   -
tcp        0      0 127.0.0.1:8332          127.0.0.1:15063         TIME_WAIT   -
tcp        0      0 127.0.0.1:8332          127.0.0.1:14371         TIME_WAIT   -
tcp        0      0 127.0.0.1:8332          127.0.0.1:14618         TIME_WAIT   -
tcp        0      0 127.0.0.1:8332          127.0.0.1:13959         TIME_WAIT   -
tcp        0      0 127.0.0.1:8332          127.0.0.1:14995         TIME_WAIT   -
tcp        0      0 127.0.0.1:8332          127.0.0.1:14513         TIME_WAIT   -
tcp        0      0 127.0.0.1:8332          127.0.0.1:14593         TIME_WAIT   -
tcp        0      0 127.0.0.1:8332          127.0.0.1:14952         TIME_WAIT   -
tcp        0      0 127.0.0.1:8332          127.0.0.1:14322         TIME_WAIT   -
tcp        0      0 127.0.0.1:8332          127.0.0.1:14563         TIME_WAIT   -
tcp        0      0 127.0.0.1:8332          127.0.0.1:14355         TIME_WAIT   -

root@bitcoins:/usr/src/pushpool# netstat -anp | grep 8332 | wc -l
1508

[xf2_org]

See comments in the other thread by BTCguild pool operator (elethuria(sp??)).

[Jine]

Going through his posts as we speak, could you point me in any more specific direction?
Only thing i can thinks of is amount of opened files (ulimit -n)

[MyFarm]

Has anyone been able to get pushpool working on an CentOS system?

[FreddyE]

Is it possibly to compile this for windows? I got visual studio but how would I go about all those dependencies?

[error]

Has anyone been able to get pushpool working on an CentOS system?

Yes. Are you having a specific issue?

Is it possibly to compile this for windows? I got visual studio but how would I go about all those dependencies?

No. Even if you could compile it, you don't even want to THINK about running anything serious like this on Windows.

[davout]

Is it possibly to compile this for windows? I got visual studio but how would I go about all those dependencies?

Why, do you want to run it on your windows mobile phone ?

[FreddyE]

No, but I have a webserver running on windows.

[FreddyE]

No. Even if you could compile it, you don't even want to THINK about running anything serious like this on Windows.

What a strange thing to say...
Who says you can´t run anything serious on Windows?

[davout]

What a strange thing to say...
Who says you can´t run anything serious on Windows?

You can successfully run serious stuff on windows, but if you try to run open source software that was built on and runs on linux you're on your own, nobody wants to deal with the CLR... Do you hear a lot from the Mono project ? I don't

I managed to install a full Ruby on Rails stack on windows, but god that was painful at some times, especially when you need some special extension that automagically gets compiled on linux but not on windows...

[FreddyE]

Yeah....manually clearing all dependencies is the issue. Oh well...seems I´m just going to install Ubuntu on a second partition.

[CubedRoot]

Right now I have my own IBM xSeries 346 with 8 gig ram and dual Xeon 3.8's running on a raid 5  with 6 scsi ultra 320's (15k RPM) array that has almost 750 gig available storage.  I have it colocated in a nice datacenter and have 13 IP's available for use on it.  It is connected to the network by 2 100meg circuits, which I have nic bonding setup for.

I am running Centos 5.6 on it right now, and I also run DirectAdmin on it and host about 38 sites on the machine. I currently have almost no load on the box with these accounts.  I also have 1.5 TB of transfer free in my SLA with the datacenter.

I am wanting to start a semi private pool for me and a few friends and possibly some more folks later on. I have been looking at pushpool for the backend, and a good friend of mine is a damn good php and sql programmer by trade.  He has agreed to do the webfrontend and DB side of things.  Is there any documentation on setting up pushpool on Centos 5.6 that I can read up on to learn about it?  

[Jine]

I'm having a hard time figuring out why my bitcoind+pushpoold setup is resulting in about 5-6% stale/duplicated.
Any ideas - anyone?!

[dikidera]

I'm having a hard time figuring out why my bitcoind+pushpoold setup is resulting in about 5-6% stale/duplicated.
Any ideas - anyone?!

You said "Finally a usable bitcoin pool" Seems that may not be the case.

[davout]

You said "Finally a usable bitcoin pool" Seems that may not be the case.

The intended audience are developers, and it's perfectly usable for developers

[genewitch]

Pushpool on gentoo won't compile, make log below:

http://paste.pocoo.org/show/404358/

damn, remove --as-needed, run make clean && make, let it fail copy the line that says gcc and take out the ,-O1 -W1, part. the commas were breaking the command. is this gentoo's make's fault or some other script that is breaking?


gcc  -g -O2 -pthread -Wl,-O1 -Wl,-rdynamic -L/usr/lib64/mysql -lmysqlclient -L/usr/lib64 -lz -lcrypt -lnsl -lm -L/usr/lib64/ -lssl -lcrypto  -o pushpoold hist.o htab.o anet.o config.o msg.o server.o util.o db-sqlite.o db-mysql.o db-postgresql.o -lcurl -lssl -lcrypto -lrt -lssl -lcrypto -ldl -lz -lz -levent -lpthread -ljansson -lcrypto -lz -lmemcached

becomes
 gcc  -g -O2 -pthread -Wl -rdynamic -L/usr/lib64/mysql -lmysqlclient -L/usr/lib64 -lz -lcrypt -lnsl -lm -L/usr/lib64/ -lssl -lcrypto  -o pushpoold hist.o htab.o anet.o config.o msg.o server.o util.o db-sqlite.o db-mysql.o db-postgresql.o -lcurl -lssl -lcrypto -lrt -lssl -lcrypto -ldl -lz -lz -levent -lpthread -ljansson -lcrypto -lz -lmemcached

testing it now.

[itsagas]

what is the best way to scale pushpool and make it distributed. 

for example, multiple pushpools and multiple bitcoind?  multiple pushpools and one bitcoind?  other? 

if for example we wanted one server in EU, one in USA and the users to share the work and the pool acts like one big pool with common accounts / workers database.

[kjj]

Run bitcoind on each server, alongside pushpoold.