[CLOSED] BTC Guild - Pays TxFees+NMC, Stratum, VarDiff, Private Servers

[bulanula]

Added Pool Block History page, API for it will be coming soon as well so web services can scrape our block data.

Regarding bitlane's post, yes, there was about 15 seconds of downtime last night for a pool restart.

MM coming soon I assume then ?

Not until shadders releases the final PSJ which includes built in work generation for merged mining.

So how come other pools already have MM ? I am not too well informed on MM.

[dmcurser]

please go back and read through the pages mm isnt really totaly stable and namce coin isnt worth that much atm and it creates alot more stalls on the bitcoin side e has a post of why he hasnt implmented it yet i think on page 4.

[eleuthria]

Added Pool Block History page, API for it will be coming soon as well so web services can scrape our block data.

Regarding bitlane's post, yes, there was about 15 seconds of downtime last night for a pool restart.

MM coming soon I assume then ?

Not until shadders releases the final PSJ which includes built in work generation for merged mining.

So how come other pools already have MM ? I am not too well informed on MM.

I stated earlier:  I wont' implement MM until it has a close to 0 effect on BTC mining, regardless of whether the tradeoff is 1% loss of BTC for a 5-7% gain in profit due to NMC.  Shadders is implementing a change to PSJ soon where the getwork creation is prepared within PSJ, rather than relying on bitcoind which has always been the bottleneck at longpoll time.  That change should reduce the performance impact of merged mining enough to be worth implementing.

[eleuthria]

Updated Auto Payout rules slightly, you can now enter multiples of 0.1 instead of whole bitcoins.

[bulanula]

Added Pool Block History page, API for it will be coming soon as well so web services can scrape our block data.

Regarding bitlane's post, yes, there was about 15 seconds of downtime last night for a pool restart.

MM coming soon I assume then ?

Not until shadders releases the final PSJ which includes built in work generation for merged mining.

So how come other pools already have MM ? I am not too well informed on MM.

I stated earlier:  I wont' implement MM until it has a close to 0 effect on BTC mining, regardless of whether the tradeoff is 1% loss of BTC for a 5-7% gain in profit due to NMC.  Shadders is implementing a change to PSJ soon where the getwork creation is prepared within PSJ, rather than relying on bitcoind which has always been the bottleneck at longpoll time.  That change should reduce the performance impact of merged mining enough to be worth implementing.

So then are you saying that pools currently using MM are in fact making us get less BTC in total than before even with MM ? Unlikely. However I think others like Eligius already have custom MM code which isolates the BTC part from NMC so that is protected in the end and not affected etc. PoolServerJ seems like a generic solution to me.

[eleuthria]

Added Pool Block History page, API for it will be coming soon as well so web services can scrape our block data.

Regarding bitlane's post, yes, there was about 15 seconds of downtime last night for a pool restart.

MM coming soon I assume then ?

Not until shadders releases the final PSJ which includes built in work generation for merged mining.

So how come other pools already have MM ? I am not too well informed on MM.

I stated earlier:  I wont' implement MM until it has a close to 0 effect on BTC mining, regardless of whether the tradeoff is 1% loss of BTC for a 5-7% gain in profit due to NMC.  Shadders is implementing a change to PSJ soon where the getwork creation is prepared within PSJ, rather than relying on bitcoind which has always been the bottleneck at longpoll time.  That change should reduce the performance impact of merged mining enough to be worth implementing.

So then are you saying that pools currently using MM are in fact making us get less BTC in total than before even with MM ? Unlikely. However I think others like Eligius already have custom MM code which isolates the BTC part from NMC so that is protected in the end and not affected etc. PoolServerJ seems like a generic solution to me.

Merged mining requires extra long polls when the NMC block changes.  Everytime you have a LP, you increase the odds of users getting stales during that window where the pool is generating thousands of getworks for miners.  MM more than doubles the number of long polls.  The only way around that FACT, is if some of the other MM pools are not pushing out a longpoll when the NMC block changes.  If that's the case, they're damaging NMC production without hurting BTC much (there is always some minor overhead in generating the work for a merged block vs a regular block).

Yes, it may be a poor excuse to some, but I've not hidden the fact that I'm not a supporter of merged mining in the first place.

[bulanula]

Added Pool Block History page, API for it will be coming soon as well so web services can scrape our block data.

Regarding bitlane's post, yes, there was about 15 seconds of downtime last night for a pool restart.

MM coming soon I assume then ?

Not until shadders releases the final PSJ which includes built in work generation for merged mining.

So how come other pools already have MM ? I am not too well informed on MM.

I stated earlier:  I wont' implement MM until it has a close to 0 effect on BTC mining, regardless of whether the tradeoff is 1% loss of BTC for a 5-7% gain in profit due to NMC.  Shadders is implementing a change to PSJ soon where the getwork creation is prepared within PSJ, rather than relying on bitcoind which has always been the bottleneck at longpoll time.  That change should reduce the performance impact of merged mining enough to be worth implementing.

So then are you saying that pools currently using MM are in fact making us get less BTC in total than before even with MM ? Unlikely. However I think others like Eligius already have custom MM code which isolates the BTC part from NMC so that is protected in the end and not affected etc. PoolServerJ seems like a generic solution to me.

Merged mining requires extra long polls when the NMC block changes.  Everytime you have a LP, you increase the odds of users getting stales during that window where the pool is generating thousands of getworks for miners.  MM more than doubles the number of long polls.  The only way around that FACT, is if some of the other MM pools are not pushing out a longpoll when the NMC block changes.  If that's the case, they're damaging NMC production without hurting BTC much (there is always some minor overhead in generating the work for a merged block vs a regular block).

Yes, it may be a poor excuse to some, but I've not hidden the fact that I'm not a supporter of merged mining in the first place.

Very nice explanation. So then this explains why Eligius is giving me less NMC but more BTC and slush is giving me a ton of NMC but less BTC Thanks !

[eleuthria]

Fixed a bug in the auto payouts.  Made a change after the first round of testing to make sure it wasn't possible to send a payment to somebody without logging it with a batch # so I could audit the results in case a batch failed.  One of those changes ended up making the payment script stop after the first payment was processed, so only 1 person would get an auto payout per hour.  That has been fixed and now the auto payouts are working as intended.

[sadpandatech]

  Pool hash way down and I can't keep connection.

  Working on something or is there something wrong?

[slush]

Merged mining requires extra long polls when the NMC block changes.  Everytime you have a LP, you increase the odds of users getting stales during that window where the pool is generating thousands of getworks for miners.  MM more than doubles the number of long polls.  The only way around that FACT, is if some of the other MM pools are not pushing out a longpoll when the NMC block changes.  If that's the case, they're damaging NMC production without hurting BTC much (there is always some minor overhead in generating the work for a merged block vs a regular block).

You're right that longpolling on NMC blocks isn't worth of doing, I'm not triggering LP on NMC block, too. However I must disagree that merged mining is adding any measurable additional load on servers when doing properly (so - without merged mining proxy) or hurt bitcoin mining in any way. Also that drop in NMC performance because of not doing LP is really mininmal, in few percents, so that's nothing what really hurt anybody.

Otherwise I agree, introducing merged mining WAS pain and I really understand that you don't want to play with it when NMC price is so low (actually it's on 30% of price in time of MM started).

[bulanula]

Merged mining requires extra long polls when the NMC block changes.  Everytime you have a LP, you increase the odds of users getting stales during that window where the pool is generating thousands of getworks for miners.  MM more than doubles the number of long polls.  The only way around that FACT, is if some of the other MM pools are not pushing out a longpoll when the NMC block changes.  If that's the case, they're damaging NMC production without hurting BTC much (there is always some minor overhead in generating the work for a merged block vs a regular block).

You're right that longpolling on NMC blocks isn't worth of doing, I'm not triggering LP on NMC block, too. However I must disagree that merged mining is adding any measurable additional load on servers when doing properly (so - without merged mining proxy) or hurt bitcoin mining in any way. Also that drop in NMC performance because of not doing LP is really mininmal, in few percents, so that's nothing what really hurt anybody.

Otherwise I agree, introducing merged mining WAS pain and I really understand that you don't want to play with it when NMC price is so low (actually it's on 30% of price in time of MM started).

So how come your pool gets me more NMC than Eligius pool Different implementation or what.

[slush]

So how come your pool gets me more NMC than Eligius pool Different implementation or what.

Well, this is highly offtopic and I don't want to hijack eleuthria's thread, but short (and last) response. It's simply because both pools are mining namecoins with all their hashpower, but they spread namecoins only for people who claimed their interest in namecoins by providing nmc wallet. As far as Eligius members are mostly geeks (well, it's a compliment), their interest in nmc is higher than interest of members on my pool. Which means that every nmc miner on my pool get more coins, because majority of pool members don't care or don't understand what namecoin is. When all pool members claim their namecoin address, average namecoin reward should be same or similar as on Eligius.

[Olly_K]

has the pool crashed ?

[what@3]

down for me too

[bulanula]

So how come your pool gets me more NMC than Eligius pool Different implementation or what.

Well, this is highly offtopic and I don't want to hijack eleuthria's thread, but short (and last) response. It's simply because both pools are mining namecoins with all their hashpower, but they spread namecoins only for people who claimed their interest in namecoins by providing nmc wallet. As far as Eligius members are mostly geeks (well, it's a compliment), their interest in nmc is higher than interest of members on my pool. Which means that every nmc miner on my pool get more coins, because majority of pool members don't care or don't understand what namecoin is. When all pool members claim their namecoin address, average namecoin reward should be same or similar as on Eligius.

OK this makes sense. Thank you.

[os2sam]

down for me too

Me three

[sadpandatech]

down for me too

Me three

  Still down for me too. I see some hash connected, either some working through tor or just ones the anti-ddos hasn't slapped yet.  Which is likely our issue here that the ddos alarm got set off erroneously and slapped us all down. Ping times and traceroute report perfectly normal. So, either false ddos detection or the pool software itself borked.

[eleuthria]

We're back up, sorry for the downtime (always hits right as you go to bed).  There was a rogue script running major logging due to the PoolServerJ flaw we found a few days ago, just to make sure the patches were working properly.  The script didn't have a check to see if it was already running from a previous cronjob.  Looks like sometime last night the script too long and ended up running twice simultaneously, which caused it to eventually get into a death spiral and lock up.

As sadpandatech mentioned, the slowdown also managed to trigger some software based DDoS protection which is why about 75% of the pool was unable to mine.

[sadpandatech]

We're back up, sorry for the downtime (always hits right as you go to bed).  There was a rogue script running major logging due to the PoolServerJ flaw we found a few days ago, just to make sure the patches were working properly.  The script didn't have a check to see if it was already running from a previous cronjob.  Looks like sometime last night the script too long and ended up running twice simultaneously, which caused it to eventually get into a death spiral and lock up.

  =)

[bulanula]

Always like your quick response times mate. How do you do it ?

I wake up regularly during the night to check on my miners at about 2 hours interval