I've not tried on .7. I had to move on to other software quite awhile ago when I could no loner get Pushpool to scale past about 400 GH/s, and even that was with multiple servers running.
It's entirely possible the new .7 release would solve some of the problems, but the lack of threading for database access is a real problem for large scale operations.
Topic: pushpool - open source pool software - page 2. (Read 120205 times)
Neither, it just simply can't scale and gets behind and eventually fails completely.
Do you have any more technical details?
Does this apply with the 0.7 bitcoind upstream, with vastly revamped RPC?
Neither, it just simply can't scale and gets behind and eventually fails completely.
I see. So does it require increasingly more space as workers are added to the pool or is that mostly by the size of the blockchain computed so far?
Software bounds and back end share storage.
Thanks.
I am not interested in hashing power (yet), but more in the number of workers that it can handle and how it can scale against with their increase. What would the bottleneck be? Software bounds, ram, etc... ?
I am not interested in hashing power (yet), but more in the number of workers that it can handle and how it can scale against with their increase. What would the bottleneck be? Software bounds, ram, etc... ?
It will not scale past a few hundred GH/s.
What are the performance limits of PushPool?
In particular: How many concurrent workers can be connected concurrently? And how much bandwidth/disk space/ram is needed per worker?
In particular: How many concurrent workers can be connected concurrently? And how much bandwidth/disk space/ram is needed per worker?
Absolutely, the diff can be changed. Just a couple constants, and an additional check.
Is there a way to configure the difficulty of the work units issued to miners?
My understanding (please correct me if wrong) that pretty much all mining pools (exception of p2pool) use a difficulty of 1 for work units. I'm concerned with the ASICs coming down the line that will be too many work units and create trouble for mining pools and miners.
With my two BFL singles I see a work unit every few seconds. With an ASIC 40GH/s single I would expect to see many per second which seems to be unnecessary network chatter and unnecessary DB work for the pool.
My understanding (please correct me if wrong) that pretty much all mining pools (exception of p2pool) use a difficulty of 1 for work units. I'm concerned with the ASICs coming down the line that will be too many work units and create trouble for mining pools and miners.
With my two BFL singles I see a work unit every few seconds. With an ASIC 40GH/s single I would expect to see many per second which seems to be unnecessary network chatter and unnecessary DB work for the pool.
I have created a pool http://alvarez.sfek.kz based on pushpool. Everything works fine but for some reason a lot of prevhash-stale and duplicate stale errors. How can I fix these errors to have been less? The rate of the pool until 2 may be necessary to add memory or remove support for memcached?
Code:
gcc -DHAVE_CONFIG_H -I. -fno-strict-aliasing -pthread -g -O2 -MT server.o -MD -MP -MF .deps/server.Tpo -c -o server.o server.c
server.c:106:2: erro: #error ("No valid database engines defined")
make[1]: ** [server.o] Error 1
make[1]: Saindo do diretório `/home/****/Pool'
make: ** [all] Erro 2
server.c:106:2: erro: #error ("No valid database engines defined")
make[1]: ** [server.o] Error 1
make[1]: Saindo do diretório `/home/****/Pool'
make: ** [all] Erro 2
I follow all the steps. But keep returning this error. What can i do?
You need to install postgresql, mysql and/or sqlite development libraries.
Code:
gcc -DHAVE_CONFIG_H -I. -fno-strict-aliasing -pthread -g -O2 -MT server.o -MD -MP -MF .deps/server.Tpo -c -o server.o server.c
server.c:106:2: erro: #error ("No valid database engines defined")
make[1]: ** [server.o] Error 1
make[1]: Saindo do diretório `/home/****/Pool'
make: ** [all] Erro 2
server.c:106:2: erro: #error ("No valid database engines defined")
make[1]: ** [server.o] Error 1
make[1]: Saindo do diretório `/home/****/Pool'
make: ** [all] Erro 2
I follow all the steps. But keep returning this error. What can i do?
Trying to fire up a pool on Ubuntu Server 10.04.3. Using libmysqld-dev. configure and make run fine, but I can't start pushpoold. There are no outputs.
How to get it running?
How to get it running?
does it support the new bitcoin version ? 0.4 ?
Yes. 0.4 does not introduce any breaking changes for pushpool.
does it support the new bitcoin version ? 0.4 ?
yeah I had to update libevent, from what I understand ./configure shuold of failed though.
anyone having trouble compiling the newest source? Ive compiled pushpool before but im getting the following:
Code:
/server.c:742: undefined reference to `evbuffer_get_length'
/pushpool/server.c:744: undefined reference to `evbuffer_pullup'
collect2: ld returned 1 exit status
make[1]: *** [pushpoold] Error 1
/pushpool/server.c:744: undefined reference to `evbuffer_pullup'
collect2: ld returned 1 exit status
make[1]: *** [pushpoold] Error 1
No problems compiling here... you may need to upgrade your libevent version and/or make sure you have the libevent-devel package installed.
anyone having trouble compiling the newest source? Ive compiled pushpool before but im getting the following:
Code:
/server.c:742: undefined reference to `evbuffer_get_length'
/pushpool/server.c:744: undefined reference to `evbuffer_pullup'
collect2: ld returned 1 exit status
make[1]: *** [pushpoold] Error 1
/pushpool/server.c:744: undefined reference to `evbuffer_pullup'
collect2: ld returned 1 exit status
make[1]: *** [pushpoold] Error 1
For those on centos, downloading sqlite, and doing ./configure --prefix=/usr/local then make/make install got me one step to getting the thing running.
stuck on this now
http://pastebin.com/7D5JL3fY
stuck on this now
http://pastebin.com/7D5JL3fY
Jump to: