I added another instance on port 9007 as a test. If some people want to try to connect to that port to see if it clears up any problems and report back, that'd be great.
Trying it for the last 10min, working fine for me.
Topic: [1200 TH] EMC: 0 Fee DGM. Anonymous PPS. US & EU servers. No Registration! - page 178. (Read 499709 times)
October 23, 2011, 05:50:43 PM
Trying it for the last 10min, working fine for me.
October 23, 2011, 05:33:43 PM
I added another instance on port 9007 as a test. If some people want to try to connect to that port to see if it clears up any problems and report back, that'd be great.
October 23, 2011, 05:04:33 PM
I'll instrument the code I'm running to check. One thing that might help is just increasing the number of front ends you're running, even on the same systems.
I haven't hacked on pushpool for a while, but when I last did it had a lot of blocking IO.... so a few slow sockets could really bog down its response time. I actually have no clue what you're running but if has a similar design you might get better throughput just by running more processes. (You could load balance with some OS level rounding-robbining of the sockets or just by giving users more ports to connect to)
I haven't hacked on pushpool for a while, but when I last did it had a lot of blocking IO.... so a few slow sockets could really bog down its response time. I actually have no clue what you're running but if has a similar design you might get better throughput just by running more processes. (You could load balance with some OS level rounding-robbining of the sockets or just by giving users more ports to connect to)
I'll look into that - I hope to have a Poolserverj test instance up this week, so it may fix the problems all together anyway.
Just joined your pool, connected to the pacific server and was getting 15% rejects, changed to the us server and haven't had a reject yet. Something wrong with the pacific server?
Yeah, there's some routing issues with that. Use the US server for the time being, it's much more reliable. I need to find a new host in Australia or Asia somewhere but so far everywhere I've looked wants a ridiculous amount of money for a colocated server.
Will you add a feature to automatically convert the NMC balance to BTC?
You could make a daily job that sends the NMC of everyone with the option selected to an exchange, withdraws, and distribute the earned BTC in proportion to the namecoins taken.
You could make a daily job that sends the NMC of everyone with the option selected to an exchange, withdraws, and distribute the earned BTC in proportion to the namecoins taken.
Is there an exchange that has a permanent API that will allow me to deposit money and initiate sales? Bitparking doesn't seem to have an API at all.
October 23, 2011, 08:35:46 AM
Will you add a feature to automatically convert the NMC balance to BTC?
You could make a daily job that sends the NMC of everyone with the option selected to an exchange, withdraws, and distribute the earned BTC in proportion to the namecoins taken.
You could make a daily job that sends the NMC of everyone with the option selected to an exchange, withdraws, and distribute the earned BTC in proportion to the namecoins taken.
October 23, 2011, 07:04:47 AM
Just joined your pool, connected to the pacific server and was getting 15% rejects, changed to the us server and haven't had a reject yet. Something wrong with the pacific server?
October 23, 2011, 03:28:28 AM
During an LP, the system is trying to hand out a ton of work at the same time, so there is an occasional bottleneck there.
I'll instrument the code I'm running to check. One thing that might help is just increasing the number of front ends you're running, even on the same systems.
I haven't hacked on pushpool for a while, but when I last did it had a lot of blocking IO.... so a few slow sockets could really bog down its response time. I actually have no clue what you're running but if has a similar design you might get better throughput just by running more processes. (You could load balance with some OS level rounding-robbining of the sockets or just by giving users more ports to connect to)
October 22, 2011, 02:30:23 PM
Yes, roll.ntime is enabled, or should be at any rate.
I am looking at switching the getwork server over to PoolServerJ... I just haven't had time to setup my java development environment and make the changes necessary. I hope to get to that soon, and assuming it's everything everyone says it is, that will hopefully correct any problems, if they exist... though I can't figure out where they might be, unless the miner is running dry during an LP.
During an LP, the system is trying to hand out a ton of work at the same time, so there is an occasional bottleneck there.
The hashrate will vary widely on the pool as to what it's reporting because it's only an estimate based on the amount of shares you've submitted, whereas the miner knows exactly what rate it's processing - so the miner hashrate will be fairly constant, but the pool hashrate will vary widely depending on your luck. Mine goes from anywhere between ~12.1 - ~14 GH/s at any given time, depending on luck.
I am looking at switching the getwork server over to PoolServerJ... I just haven't had time to setup my java development environment and make the changes necessary. I hope to get to that soon, and assuming it's everything everyone says it is, that will hopefully correct any problems, if they exist... though I can't figure out where they might be, unless the miner is running dry during an LP.
During an LP, the system is trying to hand out a ton of work at the same time, so there is an occasional bottleneck there.
The hashrate will vary widely on the pool as to what it's reporting because it's only an estimate based on the amount of shares you've submitted, whereas the miner knows exactly what rate it's processing - so the miner hashrate will be fairly constant, but the pool hashrate will vary widely depending on your luck. Mine goes from anywhere between ~12.1 - ~14 GH/s at any given time, depending on luck.
October 22, 2011, 10:45:31 AM
I'm seeing the same issue with cgminer on fast systems— watching the traffic I'm seeing the pool take several seconds to respond to getwork requests and the miner runs dry. I _assume_ this is hurting my hashrate as you're reporting 1.80GH/s against a system which does 1940.6 MH/s day in and out [...]
I observe something similar on my phoenix miners, I can see them waiting and dropping to 0 hashrate. And on EMC my hashrate report varies between 2.6 and 2.9 GH/s, meaning it slows down although my miners are running at constant speed. 
October 22, 2011, 10:30:48 AM
I think the issue might be Nvidia-related, because my two ATI rigs don't seem to be affected. Switching from cgminer v2.0.3 to v2.0.4 seems to have helped, too.
Now, shares-per-minute has gone up for all my workers. Life is good. Awesome pool you've got here, by the way.
Now, shares-per-minute has gone up for all my workers. Life is good. Awesome pool you've got here, by the way.
October 22, 2011, 07:54:26 AM
Is your hashrate/expected shares suffering at all, Badpenny?
Everything seems normal otherwise. My hashrate seems correct, share count is about right, maybe a little low. Current sample shows 1066 shares submitted with 3 stale, so that's good. Average shares per minute is about what it is when mining at Arsbitcoin or Mt. Red.
October 22, 2011, 12:47:26 AM
I've seen that error from cgiminer on occasion and never been able to trace it back to anything on the pool. I've chalked it up to either a bug or overzealous timing in cgiminer, as poclbm and phoenix aren't having any issues and reject rate is running about .2%. (Yes, .2%, not 2%)
Is anyone else experiencing any issues of any kind?
Is anyone else experiencing any issues of any kind?
I'm seeing the same issue with cgminer on fast systems— watching the traffic I'm seeing the pool take several seconds to respond to getwork requests and the miner runs dry. I _assume_ this is hurting my hashrate as you're reporting 1.80GH/s against a system which does 1940.6 MH/s day in and out, though maybe I've just had bad timing. I do see the 0.2% stale rate, which is nice enough.
Cgminer systems can have a lot of hash rate behind a single connection, and they avoid queuing much work to avoid stales... so it's important to keep them saturated.
I didn't bother reconnecting to check the headers, do you have nrolltime enabled?
October 22, 2011, 12:31:22 AM
I've seen that error from cgiminer on occasion and never been able to trace it back to anything on the pool. I've chalked it up to either a bug or overzealous timing in cgiminer, as poclbm and phoenix aren't having any issues and reject rate is running about .2%. (Yes, .2%, not 2%)
Is anyone else experiencing any issues of any kind? I have some tweaks I can make to speed things up if need be, but I've held off because I don't want to introduce anything that might cause some initial problems into a system that seems to be working fine.
Is your hashrate/expected shares suffering at all, Badpenny?
Is anyone else experiencing any issues of any kind? I have some tweaks I can make to speed things up if need be, but I've held off because I don't want to introduce anything that might cause some initial problems into a system that seems to be working fine.
Is your hashrate/expected shares suffering at all, Badpenny?
October 21, 2011, 10:28:13 PM
I keep getting "Pool 0 not providing work fast enough" from cgminer @ Eclipse, each one exactly one minute apart. 
October 21, 2011, 02:19:53 PM
Yeah... we are? How's that? Did we lose Slush and Deepbit and BTCGuild and stuff?
Our hash rate is also slightly increased compared to the past weeks. Maybe because you implemented merged mining? October 20, 2011, 11:24:37 PM
You can ignore those options. The scored pool is separate from the prop pool now and the DGS basically takes the place of LNSPPS.
I should probably remove that drop down.
I should probably remove that drop down.
October 20, 2011, 08:17:12 PM
Hi there, thanks for the pool!
I was wondering, when I create a worker I can choose between Score/Prop. and LNSPPS, I have a hard time figuring out which one to choose, could someone explain the differences between those two for me please?
Thanks in advance.
I was wondering, when I create a worker I can choose between Score/Prop. and LNSPPS, I have a hard time figuring out which one to choose, could someone explain the differences between those two for me please?
Thanks in advance.
October 20, 2011, 07:48:01 PM
Yeah... we are? How's that? Did we lose Slush and Deepbit and BTCGuild and stuff?
October 20, 2011, 06:28:01 PM
You guys are stickied for the first time I think?
October 20, 2011, 02:15:54 PM
I made some changes so the check boxes should work properly now.
October 20, 2011, 11:17:34 AM
Early on, I received some texts for NMC even though it wasn't checked(by default).
I checked and saved it, then unchecked and saved again.
Haven't received any since.
I checked and saved it, then unchecked and saved again.
Haven't received any since.
Jump to: