█▓▒░-< [ZPOOL.CA][BTC Multipool] The miners multipool >-░▒▓█ Paid 925+ BTC

[JaredKaragen]

My workers get dead hashrait 15-35%, as from sha-256 - algo, and  scrypt - algo too! (on the zpool .ca)
I need your help!

try adjusting your diff?

only very recently for the first time have I ever seen rejected hashrate from zpool in the past 2-3 days.

And that's from machines on a nice connection in SF....  But I still haven't see them on my main machine at home on the cellular link.  So Id chalk it personally up to the connection you are using... a node down the line or something.

If you get your difficulty set right;  your hashrate on the site will more or less be within 10% of your miner at any given time.

As far as SHA256 disconnects;  your hashrate average drops to zero and must build up again, so the graph will not be exactly what your hashrates actually were.
You may notice your hashrate drop a ton or completely;  look at the pool statistics and notice that the hashrate probably dropped substantially and is building back up again....  look at the pool statistics for sha256;  you see the spikes in hashrates/profitability lined up as they happen.  sometimes they are more frequent than others.   Faster miners take a harder hit than slower miners with this issue.  I have no issues keeping my S3's pointed at zpool;  but I just have my s7 on solo bitcoin.   Too much intervention needed from the disconnects causing issues with the miner switching pools, etc.

[smilodon]

bug to c11 algo maximum shares 79.3800%

[AngryDwarf]

bug to c11 algo maximum shares 79.3800%

Long discussed outstanding issue that effects all algos.
Even more scary is the 5% sha256 fee and whopping 20% scrypt fee!!! Surely a development system to production system rollout error.

[crackfoo]

bug to c11 algo maximum shares 79.3800%

Even more scary is the 5% sha256 fee and whopping 20% scrypt fee!!! Surely a development system to production system rollout error.

Whoops, didn't notice that! I forked the pool software to the next version and missed the example where you can set per algo fees.


https://github.com/tpruvot/yiimp/blob/next/web/serverconfig.sample.php#L73

[jkminkov]

bug to c11 algo maximum shares 79.3800%

myr-gr is ok, I did have 90-something % for a few hours.

Reject* however shows some strange high percentile. miner side shows from 0 to 2,1% reject.

[joblo]

Nice to see some new algos in the pool.

Is blakecoin showing the correct hashrate? Wrong units maybe?

[crackfoo]

Nice to see some new algos in the pool.

Is blakecoin showing the correct hashrate? Wrong units maybe?

Yeah the units doesn't seem to be right. Working on that.

[Megaquake]

I like the development of this pool but there is other issues that need to be addressed like the stratum for SHA256. not only does it disconnect but its a miner killer, it sits their under full load not hashing forever unless you you do a hard reset after a short period, so many complaints about it with no resolution, all you have done is say check your diff setting, its not the miner its the pool.

[AngryDwarf]

I like the development of this pool but there is other issues that need to be addressed like the stratum for SHA256. not only does it disconnect but its a miner killer, it sits their under full load not hashing forever unless you you do a hard reset after a short period, so many complaints about it with no resolution, all you have done is say check your diff setting, its not the miner its the pool.

It does happen on other ports, but no where near as frequently as the sha256 port. There are a lot of fast coins on that port.
I've tried running the software myself with limited success, I tried getting the qubit stratum running and that was constantly resetting. Running a pool such as this requires a lot of RAM and good disk speed, and only a handful of wallets pushes my 16GB PC into swap territory. Looking at the hardware that zpool runs, I'd say it's time to bite the bullet and go to SSD disks.

[AngryDwarf]

Is there any reason not to add skein, the 5th Algo for DGB/MYR/AUR?

[crackfoo]

Is there any reason not to add skein, the 5th Algo for DGB/MYR/AUR?

I'll be added a few more algos over the next few weeks.

btw, the pool is already running of SSD's and is also distributed over 4  8-12core  servers with 256GB ram. It's not the hardware. I've changed it 3 times now. The hardware investment is now 10-15k a year. I'm still looking for a dev that is willing to re-write port the stratum component to something more robust and stable. I have one prospect so far and it by no means cheap. It will be a $2-3k job and has to be done right. Trust me, nothing I hate more than seeing the stratum resets. It's also BTC out of my pocket too.

My end game is to be A) The most profitable, B) Reliable, C) Stable and D) The biggest.

I've got 3 out of 4 IMO, stability will come as soon as I can get it developed.

[JaredKaragen]

I'll comment that cheap reliable and fast are rarely used all together to describe something.  Usually it's just two of the three that you get.


You're still doing an amazing job with the pool

[Erumara]

Is there any reason not to add skein, the 5th Algo for DGB/MYR/AUR?

I'll be added a few more algos over the next few weeks.

btw, the pool is already running of SSD's and is also distributed over 4  8-12core  servers with 256GB ram. It's not the hardware. I've changed it 3 times now. The hardware investment is now 10-15k a year. I'm still looking for a dev that is willing to re-write port the stratum component to something more robust and stable. I have one prospect so far and it by no means cheap. It will be a $2-3k job and has to be done right. Trust me, nothing I hate more than seeing the stratum resets. It's also BTC out of my pocket too.

My end game is to be A) The most profitable, B) Reliable, C) Stable and D) The biggest.

I've got 3 out of 4 IMO, stability will come as soon as I can get it developed.

The resets can be irritating, I set up a stratehm proxy and it cut my issues by a huge amount. Still have to reboot the rigs every few days but they rarely lock up anymore and it's a small price to pay for this kind of profitability. Glad to hear the issues are still in the process of being solved, I can definitely understand rewriting the likely outdated protocols would be quite the complicated job.

In any case, keep up the good work crackfoo!

[Nataly2014]

ERUMARA
...I set up a stratehm proxy ....
You can provide a link to instructions for creating your proxy stratehm?
And what requirements hardware?
Thanks!

[AngryDwarf]

Is there any reason not to add skein, the 5th Algo for DGB/MYR/AUR?

I'll be added a few more algos over the next few weeks.

btw, the pool is already running of SSD's and is also distributed over 4  8-12core  servers with 256GB ram. It's not the hardware. I've changed it 3 times now. The hardware investment is now 10-15k a year. I'm still looking for a dev that is willing to re-write port the stratum component to something more robust and stable. I have one prospect so far and it by no means cheap. It will be a $2-3k job and has to be done right. Trust me, nothing I hate more than seeing the stratum resets. It's also BTC out of my pocket too.

My end game is to be A) The most profitable, B) Reliable, C) Stable and D) The biggest.

I've got 3 out of 4 IMO, stability will come as soon as I can get it developed.

Ouch. I hope you get a highly configurable stratum server for that job. It's a bit beyond something I would want to tackle at the moment.
The issue I was getting involved the monitor thread. I was trying to use it to solo mine some MYR which now has the getwork protocol removed. For some reason the system thought my MYR address was not a valid MYR address, so I think I was getting this due to the stratum server not having any jobs to provide.
If your stratums are resetting at the same piece of code (perhaps due to getblocktemplate rpc delays), than I suppose you could always try increasing the max job delay. It might mean the stratum clients aren't being fed with jobs to run at full whack at times, but it might be better than a reset.

stratum.h

Code:

#define YAAMP_MAXJOBDELAY		(2*60)

stratum.cpp

Code:

///////////////////////////////////////////////////////////////////////////////

void *monitor_thread(void *p)
{
        while(1)
        {
                sleep(120);

                if(g_last_broadcasted + YAAMP_MAXJOBDELAY < time(NULL))
                {
                        stratumlog("%s dead lock, exiting...\n", g_current_algo->name);
                        exit(1);
                }
        }
}

///////////////////////////////////////////////////////////////////////////////

[Erumara]

Resources I used to get started:

https://bitcointalk.org/index.php?topic=611338.0

https://www.nicehash.com/?p=asic

Both are full of valuable info and contain basically everything I needed to get running. I run the java proxy on my windows 7 desktop and run it through a VPN and have no issues with connectivity. It barely takes any resources (~260MB of RAM and almost no processor time) and is quite stable.
I'm looking at turning a raspberry pi into a linux stratehm server and taking the load off of my desktop, looking to get my hands on one in the next month or so and I will definitely keep you updated.

Some things to consider:
Set as many failovers as you want, I don't think there's an upper limit. And it only takes one spot on your GUI menu (antminer etc.) so you can still set hard failovers should your stratum server go down. Pool swap and change priorities in a GUI without even opening your miner settings!

If you have any specific questions regarding setup I will try to help as best I can

Tips:
I set my reconnect period to 15 seconds and the stability testing to 30. Seem to work great and catch zpool back up while most of the miners are still deciding what pool to stay connected to. Also largely fixes the issue with units locking up as they're not constantly getting a connections, sending shares, and getting disconnected again.

Watch your sampling periods, I had it set to 5 secs to start and it hung the server for a few seconds every update, now set to 300secs and works great.

Combine your stratehm server with Cryptoglance for monitoring and laugh your way to the exchange

Cheers,

Erumara

[crackfoo]

Is there any reason not to add skein, the 5th Algo for DGB/MYR/AUR?

I'll be added a few more algos over the next few weeks.

btw, the pool is already running of SSD's and is also distributed over 4  8-12core  servers with 256GB ram. It's not the hardware. I've changed it 3 times now. The hardware investment is now 10-15k a year. I'm still looking for a dev that is willing to re-write port the stratum component to something more robust and stable. I have one prospect so far and it by no means cheap. It will be a $2-3k job and has to be done right. Trust me, nothing I hate more than seeing the stratum resets. It's also BTC out of my pocket too.

My end game is to be A) The most profitable, B) Reliable, C) Stable and D) The biggest.

I've got 3 out of 4 IMO, stability will come as soon as I can get it developed.

Ouch. I hope you get a highly configurable stratum server for that job. It's a bit beyond something I would want to tackle at the moment.
The issue I was getting involved the monitor thread. I was trying to use it to solo mine some MYR which now has the getwork protocol removed. For some reason the system thought my MYR address was not a valid MYR address, so I think I was getting this due to the stratum server not having any jobs to provide.
If your stratums are resetting at the same piece of code (perhaps due to getblocktemplate rpc delays), than I suppose you could always try increasing the max job delay. It might mean the stratum clients aren't being fed with jobs to run at full whack at times, but it might be better than a reset.

stratum.h

Code:

#define YAAMP_MAXJOBDELAY		(2*60)

stratum.cpp

Code:

///////////////////////////////////////////////////////////////////////////////

void *monitor_thread(void *p)
{
        while(1)
        {
                sleep(120);

                if(g_last_broadcasted + YAAMP_MAXJOBDELAY < time(NULL))
                {
                        stratumlog("%s dead lock, exiting...\n", g_current_algo->name);
                        exit(1);
                }
        }
}

///////////////////////////////////////////////////////////////////////////////

Yup, I've tried manipulating those even up to the 600, it then just hangs that much longer.  There were a few other timeouts I tried manipulating as well with little improvement. If only it was smart enough to identify which wallet it was that was causing the issues, and simply disconnect it... it would be so much better. To reset stratum, seems like a big hammer and IMO, was a big shortcut by the orig dev.

[AngryDwarf]

Yup, I've tried manipulating those even up to the 600, it then just hangs that much longer.  There were a few other timeouts I tried manipulating as well with little improvement. If only it was smart enough to identify which wallet it was that was causing the issues, and simply disconnect it... it would be so much better. To reset stratum, seems like a big hammer and IMO, was a big shortcut by the orig dev.

I doubt you could narrow it down to any particular coins wallet, as it seems it can be any wallet at any time that might either respond immediately or take a long time to return. I assume the wallet must be doing something that prevents it from processing rpc calls at the time.

Calling GBT with no parameters can return a massive amount of data on the BTC wallet. Do you know if it is making a GBT call on BTC? Since this is not being mined directly, it might be possible to put an exclusion in for the BTC wallet.

[Nataly2014]

Erumara

thanks for the great work!
I will study your links!

[ghotiphry]

I'm afraid I need some help with profitability normalization.  I'm running several 750ti cards.  Since the normalization rates at zpool are based on AMD, it doesn't make sense for me to use them.  What nVidia miners out there use for factors, based on the hashrate for the various algorithms?  I can't seem to figure out how to calculate these values.