KanoPool since 2014 🐈 - PPLNS and Solo 0.5% fee - Worldwide - 2438 blocks

[norgan]

And a mini bounce for further scalability improvements.

Just so people know what the improvements are, if you check what's gone into the latest ckpool code you'll see changes that correspond with the pool upgrades (doesn't get more open than that does it?).

The ckpool code was not remotely limiting, but being able to see the pool working live and profiling where the CPU is used makes for excellent instant development of whatever is currently the biggest CPU user. The main connection event handler was converted from poll to epoll reducing CPU usage of that to 1/10th and the share processing workqueue was broken up into many threads (proportional to the number of CPUs on the machine) to better distribute out share processing. We're now prepared for about 10x as many workers as we were yesterday, so keep em coming...

Nice work, I noticed the diff reset and hash go to 0 but came up pretty quick. Working well now. A question, if I am in Australia and I have several miners, would I be better to send them to a local ckpool proxy then send that to ck pool in regards to latency and reduced reject/stales?

[-ck]

Nice work, I noticed the diff reset and hash go to 0 but came up pretty quick. Working well now. A question, if I am in Australia and I have several miners, would I be better to send them to a local ckpool proxy then send that to ck pool in regards to latency and reduced reject/stales?

If your internet bandwidth is often maxed out, especially upstream (such as ADSL which most Australians have) then there is a potential advantage to combining them with a local proxy. On the other hand if you're not maxing it out, you're better off leaving them all connected separately as every extra step in the chain adds its own form of latency. The server is on the west coast of USA though so it's about as close as you can get to Australia and still be in the states.

[norgan]

Nice work, I noticed the diff reset and hash go to 0 but came up pretty quick. Working well now. A question, if I am in Australia and I have several miners, would I be better to send them to a local ckpool proxy then send that to ck pool in regards to latency and reduced reject/stales?

If your internet bandwidth is often maxed out, especially upstream (such as ADSL which most Australians have) then there is a potential advantage to combining them with a local proxy. On the other hand if you're not maxing it out, you're better off leaving them all connected separately as every extra step in the chain adds its own form of latency. The server is on the west coast of USA though so it's about as close as you can get to Australia and still be in the states.

Got them on good fibre so all good. Thanks for clearing that up.

[DanielRo]

https://blockchain.info/tx/7b47a5415f69a092f5d2c252ff6bc60f34c333e52ef04cdc24adf485f9db270d

Code:

[10:25:54] <kanoi_> blockchain only showing 1 block so that's a good sign :)
[10:26:03] <QG> yoouu pew pew pew!!!

QG

Nice. 107.41%. May the next be more quickly.

[loshia]

And a mini bounce for further scalability improvements.

Just so people know what the improvements are, if you check what's gone into the latest ckpool code you'll see changes that correspond with the pool upgrades (doesn't get more open than that does it?).

The ckpool code was not remotely limiting, but being able to see the pool working live and profiling where the CPU is used makes for excellent instant development of whatever is currently the biggest CPU user. The main connection event handler was converted from poll to epoll reducing CPU usage of that to 1/10th and the share processing workqueue was broken up into many threads (proportional to the number of CPUs on the machine) to better distribute out share processing. We're now prepared for about 10x as many workers as we were yesterday, so keep em coming...

Nice work as always
Could you be so kind you or kano to share total mem usage of the pool at the momet?
thank you

[bgibso01]

Is the worker count correct for the mining to address link?  Mine is currently showing 6 workers, but I really only have 5.  Maybe someone is being kind, except the hashrate looks correct.

[TheMinex]

Would be nice to have - min payout setting ( obviously must be higher than pool min payout  ), near dust transactions are disliked by small mines too
Mined with address for little time, then registered and I continued to use that address. Don't know if addiction of the founds is automated, keep an eye at the next block!

Cheers, great work 

[DanielRo]

Question. How to withdrawal the funds?
I've registred an user, mined for about two days. A block was found. Now it is matured.
I've entered one of my BTC adresses in Account Settings, but no withdrawal button. I may guess it is done manually at this stage of pool development. May I need to send a PM with my username?

[-ck]

Is the worker count correct for the mining to address link?  Mine is currently showing 6 workers, but I really only have 5.  Maybe someone is being kind, except the hashrate looks correct.

The worker count is correct. If you issued a "restart" on cgminer et. al then the socket may still be around from the old instance. It's something I'm fixing now on cgminer now that I'm aware it happens.

[-ck]

Question. How to withdrawal the funds?
I've registred an user, mined for about two days. A block was found. Now it is matured.
I've entered one of my BTC adresses in Account Settings, but no withdrawal button. I may guess it is done manually at this stage of pool development. May I need to send a PM with my username?

It's all manual at the moment and all payments go out together with the entire block funds. Next payment should be processed later today.

[philipma1957]

Question. How to withdrawal the funds?
I've registred an user, mined for about two days. A block was found. Now it is matured.
I've entered one of my BTC adresses in Account Settings, but no withdrawal button. I may guess it is done manually at this stage of pool development. May I need to send a PM with my username?

It's all manual at the moment and all payments go out together with the entire block funds. Next payment should be processed later today.

thanks for info. I see block has been fully confirmed so I guess payments come soon.

[DanielRo]

Question. How to withdrawal the funds?
I've registred an user, mined for about two days. A block was found. Now it is matured.
I've entered one of my BTC adresses in Account Settings, but no withdrawal button. I may guess it is done manually at this stage of pool development. May I need to send a PM with my username?

It's all manual at the moment and all payments go out together with the entire block funds. Next payment should be processed later today.

Thanks.

[bgibso01]

Is the worker count correct for the mining to address link?  Mine is currently showing 6 workers, but I really only have 5.  Maybe someone is being kind, except the hashrate looks correct.

The worker count is correct. If you issued a "restart" on cgminer et. al then the socket may still be around from the old instance. It's something I'm fixing now on cgminer now that I'm aware it happens.

Ok, that's exactly it.  I'm using some S-4 units that I have to apply your fix to.

[philipma1957]

shit did we just hit another block!

328234?

[-ck]

Yes

[philipma1957]

Yes

snap it has 5 confirms and I had all six s-3's point to it yesss!

https://blockchain.info/tx/5744a8daf67d046e486644ac3b387cd0f81e2d2497bad5d293dc37aab381ae46

life is good.

I hit me one on mmpool.org today  and one here.

[kano]

Question. How to withdrawal the funds?
I've registred an user, mined for about two days. A block was found. Now it is matured.
I've entered one of my BTC adresses in Account Settings, but no withdrawal button. I may guess it is done manually at this stage of pool development. May I need to send a PM with my username?

It's all manual at the moment and all payments go out together with the entire block funds. Next payment should be processed later today.

There was a couple of shares dropped that I've tracked down the cause and doing some changes today to allow running the "confirmer" (-y/-Y in the code) with a MUCH lower memory footprint than it currently needs.
It checks share counts - that I don't usually run unless there was a processing error at any stage during the block - which we've not had any until the previous block.
Thus once I've done that and put the shares back in I'll do the payout today.

[kano]

And a mini bounce for further scalability improvements.

Just so people know what the improvements are, if you check what's gone into the latest ckpool code you'll see changes that correspond with the pool upgrades (doesn't get more open than that does it?).

The ckpool code was not remotely limiting, but being able to see the pool working live and profiling where the CPU is used makes for excellent instant development of whatever is currently the biggest CPU user. The main connection event handler was converted from poll to epoll reducing CPU usage of that to 1/10th and the share processing workqueue was broken up into many threads (proportional to the number of CPUs on the machine) to better distribute out share processing. We're now prepared for about 10x as many workers as we were yesterday, so keep em coming...

Nice work as always
Could you be so kind you or kano to share total mem usage of the pool at the momet?
thank you

CKDB uses truck loads
CKPool is tiny.

... and an explanation of the trucks
https://bitcointalk.org/index.php?topic=790323.msg9417792#msg9417792

[philipma1957]

payout came for 1 of the two blocks.  thanks

[-ck]

Could you be so kind you or kano to share total mem usage of the pool at the momet?

CKDB uses truck loads
CKPool is tiny.

ckpool itself is using 550kb on this pool at the moment. No that's not a typo.

EDIT: 550Mb, yes it was a typo.