BAMT - Easy persistent USB key based linux for dedicated miners/mining farms

[ElectricMucus]

donations make things happen, requests without them are generally ignored.  

as I said on irc, I would be happy to if I could 

Though, after some consideration, it's not for me. Not that it isn't a nice piece of work.
But your attitude... I find it offensive. This way you pretty much defy the purpose of donations (they're voluntary) and FOSS.

Make the best of my comment... in the mean time.
Go to hell. 

Wow!

Can I ask how old you are? Please tell us.

you can ask.

Are you even involved in this project or just looking for a flamewar?

[1714211351]

1714211351

[1714211351]

1714211351

[1714211351]

1714211351

[Jazkal]

you can ask.

Are you even involved in this project or just looking for a flamewar?

Ok, since you didn't answer, I'll assume by your comments that you are 16. Until you tell me otherwise, this is the number I'll go by.

Am I involved with this project? Hmm, yes, I am an end user.

And I wasn't trying to start a flame war, I was just amazed by your comments to a developer that is working hard to provide something to the community.

[ElectricMucus]

you can believe whatever you want 
I don't have to explain myself to you. 

[gnar1ta$]

Good to see we're all getting along.  There are people here willing to donate to the developer for his time and are happy to have such an easy and efficient mining package - if that isn't you or you are no longer interested in this project maybe you could move your ramblings on to another thread.

@lodcrappo

I'm still trying to pin down the Xorg crash I'm having with extenders, and noticed the xorg-video-ati and xorg-video-radeon packages are still installed with the fglrx drivers (version?).  The install page for catalyst says to remove these, but I don't know enough about how they work.  Are the xorg packages needed with the catalyst driver or will I be ok to remove them?

[lodcrappo]

Good to see we're all getting along.  There are people here willing to donate to the developer for his time and are happy to have such an easy and efficient mining package - if that isn't you or you are no longer interested in this project maybe you could move your ramblings on to another thread.

@lodcrappo

I'm still trying to pin down the Xorg crash I'm having with extenders, and noticed the xorg-video-ati and xorg-video-radeon packages are still installed with the fglrx drivers (version?).  The install page for catalyst says to remove these, but I don't know enough about how they work.  Are the xorg packages needed with the catalyst driver or will I be ok to remove them?

Not sure... try apt-get remove package-name
If it works, great, if not, you can just 'apt-get install package-name' and should be back to how it was before.

[lodcrappo]

donations make things happen, requests without them are generally ignored.  

as I said on irc, I would be happy to if I could  

Though, after some consideration, it's not for me. Not that it isn't a nice piece of work.
But your attitude... I find it offensive. This way you pretty much defy the purpose of donations (they're voluntary) and FOSS.

Make the best of my comment... in the mean time.
Go to hell.  

So I give you software that took many hours of my time, for free.

You don't like a couple things about it.

You request I change it for you, for free.

I say no.  You must pay for my time if you want changes or support for yourself.

You say you can't afford it.

I offer to trade time, an hour of your time helping with the project for an hour of mine helping you.

You rage quit IRC.

And I'm offensive?

Fuck right off, asshole.  Nobody cares if you use BAMT.  You will do nothing positive for any of us, that is clear.

[ElectricMucus]

I offer to trade time, an hour of your time helping with the project for an hour of mine helping you.

That would be acceptable, I didn't notice that offer for whatever reason. 
Where should I start?

[lodcrappo]

I offer to trade time, an hour of your time helping with the project for an hour of mine helping you.

That would be acceptable, I didn't notice that offer for whatever reason.  
Where should I start?

[15:32:23] <ElectricMucus> Would be happy to pay it if I could :]
[15:34:59] <@Aaaron> well, you can do an hour of work for me, and i will do one for you
[15:36:37] ElectricMucus [56212f1d@gateway/web/freenode/ip.x.x.x.x] has quit IRC: Ping timeout: 252 seconds
[15:36:48] <@Aaaron> guess that wasn't an appealing offer

That offer is rescinded, for you.  Go away.

[lodcrappo]

Fix #18

Bugfix for mineco, thanks to gnar1ta$ for assistance in debugging that.  Added support for kiwi, but they only provide hash rate.

[Big Time Coin]

subscription post, going to try rolling this out to a few rigs to test.

[de4l]

Any thoughts on cgminer?

I just started using it on my windows machine and it's getting a fair amount less stales than my BAMT box.  About 0.40% stales to 1.40% stales, it's only been about 7 hours running, so it could be luck... but I think it may also be the miner.

I've been testing cgminer.  It hasn't performed significantly differently than phoenix so far for me.  BAMT's phoenix is somewhat customized for our purposes, and it would be considerable work to extend cgminer to have these features as well.  If there does end up being a clear benefit, yes we will support it or switch to it as the primary miner, but so far I'm unconvinced.

an update on cgminer...  after some more time I do like it, and it has one killer feature which is the ability to generate local work when a pool disconnects, then submit it on reconnect.  I would like to eventually replace bamt's custom phoenix with cgminer.

However, it's also buggy as hell at this time.. at least the version in git atm.  I really cannot use it in production.  Our phoenix has proven quite reliable (well.. when wrapped with bamt magic ).  I'll keep watching and testing, and hopefully at some point cgminer will be reliable enough to use in bamt.  Seems the author is actively maintaining it so this might not take long.

cgminer just had another great update today, and it looks like they're working on having the output easily parsable, which should make it easy to tie into your monitoring scripts... hate to suggest more work but cgminer would totally make BAMT a must use tool for a lot of people, the only reason I haven't switched from ubuntu 11.04 is that I'm still running cgminer on it great

[lodcrappo]

Any thoughts on cgminer?

I just started using it on my windows machine and it's getting a fair amount less stales than my BAMT box.  About 0.40% stales to 1.40% stales, it's only been about 7 hours running, so it could be luck... but I think it may also be the miner.

I've been testing cgminer.  It hasn't performed significantly differently than phoenix so far for me.  BAMT's phoenix is somewhat customized for our purposes, and it would be considerable work to extend cgminer to have these features as well.  If there does end up being a clear benefit, yes we will support it or switch to it as the primary miner, but so far I'm unconvinced.

an update on cgminer...  after some more time I do like it, and it has one killer feature which is the ability to generate local work when a pool disconnects, then submit it on reconnect.  I would like to eventually replace bamt's custom phoenix with cgminer.

However, it's also buggy as hell at this time.. at least the version in git atm.  I really cannot use it in production.  Our phoenix has proven quite reliable (well.. when wrapped with bamt magic ).  I'll keep watching and testing, and hopefully at some point cgminer will be reliable enough to use in bamt.  Seems the author is actively maintaining it so this might not take long.

cgminer just had another great update today, and it looks like they're working on having the output easily parsable, which should make it easy to tie into your monitoring scripts... hate to suggest more work but cgminer would totally make BAMT a must use tool for a lot of people, the only reason I haven't switched from ubuntu 11.04 is that I'm still running cgminer on it great

what features specifically do you want to see in BAMT from cgminer?  after playing with it, I'm not terribly interested, it doesn't do much the Phoenix lacks besides the offline work generation which is admittedly nice.  It would probably be easier to work on adding that one feature to phoenix than to hack cgminer to do all the things our custom Phoenix does.

[gnar1ta$]

Load balancing....though mostly because of my crappy ISP.

[lodcrappo]

Load balancing....though mostly because of my crappy ISP.

I'm not familiar with what they call "load balancing" in cgminer.. do you mean connecting to multiple pools/running multiple mining instances with a single GPU?  Or something fancier?

I have nearly completed a very nice proxy that will run on BAMT, allowing you to do many advanced, interesting, and silly things with how your GPUs connect to your pools.   Basically the local GPUs just all connect to a local proxy instance.  The proxy takes care of making sure the work gets done and submitted where you want it to, and nobody sits idle.  The proxy can intelligently manage connection to pools without the miner clients ever needing to worry about it... they ask for work, they get work... that's all they need to know.

IMHO, the miner client should be a simple as possible.. it should just mine.   Fancy logic doesn't belong down at that level, there should be clear separation between the calculation and the work routing layers.   Phoenix fits into this type of design nicely.

[gnar1ta$]

I'm not familiar with what they call "load balancing" in cgminer.. do you mean connecting to multiple pools/running multiple mining instances with a single GPU?  Or something fancier?

I don't know how the programming works, I set 3 workers on my main pool and 1 on a backup - not tied to a particular GPU - and all the GPU's stay working even when my connection drops to one or more, all the work is automatically shifted to working pools and back again when the connection stabalizes.  It sounds like your proxy would work the same and I much prefer the simplicity of BAMT, I just lose about 20% of my time to empty queues and reconnects with failover due to a sweet 10% packet loss and 10 fold latency jumps wich my ISP refuses to fix.

[mikeo]

...
I have nearly completed a very nice proxy that will run on BAMT, allowing you to do many advanced, interesting, and silly things with how your GPUs connect to your pools.   Basically the local GPUs just all connect to a local proxy instance.  The proxy takes care of making sure the work gets done and submitted where you want it to, and nobody sits idle.  The proxy can intelligently manage connection to pools without the miner clients ever needing to worry about it... they ask for work, they get work... that's all they need to know.

IMHO, the miner client should be a simple as possible.. it should just mine.   Fancy logic doesn't belong down at that level, there should be clear separation between the calculation and the work routing layers.   Phoenix fits into this type of design nicely.

This sounds great! And I concur with your overall strategy as it is elegant in its simplicity. Not to say it is simply done--I sure sure couldn't.

[kirax]

...
I have nearly completed a very nice proxy that will run on BAMT, allowing you to do many advanced, interesting, and silly things with how your GPUs connect to your pools.   Basically the local GPUs just all connect to a local proxy instance.  The proxy takes care of making sure the work gets done and submitted where you want it to, and nobody sits idle.  The proxy can intelligently manage connection to pools without the miner clients ever needing to worry about it... they ask for work, they get work... that's all they need to know.

IMHO, the miner client should be a simple as possible.. it should just mine.   Fancy logic doesn't belong down at that level, there should be clear separation between the calculation and the work routing layers.   Phoenix fits into this type of design nicely.

This sounds great! And I concur with your overall strategy as it is elegant in its simplicity. Not to say it is simply done--I sure sure couldn't.

Your strategy is sound, and unix-like. Commands should do one simple thing, so I can route them through pipes, grep, and awk, to do fancy things. I appreciate your hard work on this project, I am currently deploying about 5 GH with BAMT, and the setup time is the least I have ever seen, so low I'm no longer bothering to image my drives and make backups, because I have to edit all of two files in about 60 seconds. As such, there is 2.5 BTC on it's way to you shortly (waiting for confirmations to my withdrawl to avoid fees). My only possible request would be to replace or fix munin, I think I broke it when I was messing with hostnames, it is sort of a delicate flower. However, your local pool routing idea is something I have toyed with doing myself, so feel free to work on that as well!

[lodcrappo]

...
I have nearly completed a very nice proxy that will run on BAMT, allowing you to do many advanced, interesting, and silly things with how your GPUs connect to your pools.   Basically the local GPUs just all connect to a local proxy instance.  The proxy takes care of making sure the work gets done and submitted where you want it to, and nobody sits idle.  The proxy can intelligently manage connection to pools without the miner clients ever needing to worry about it... they ask for work, they get work... that's all they need to know.

IMHO, the miner client should be a simple as possible.. it should just mine.   Fancy logic doesn't belong down at that level, there should be clear separation between the calculation and the work routing layers.   Phoenix fits into this type of design nicely.

This sounds great! And I concur with your overall strategy as it is elegant in its simplicity. Not to say it is simply done--I sure sure couldn't.

Your strategy is sound, and unix-like. Commands should do one simple thing, so I can route them through pipes, grep, and awk, to do fancy things. I appreciate your hard work on this project, I am currently deploying about 5 GH with BAMT, and the setup time is the least I have ever seen, so low I'm no longer bothering to image my drives and make backups, because I have to edit all of two files in about 60 seconds. As such, there is 2.5 BTC on it's way to you shortly (waiting for confirmations to my withdrawl to avoid fees). My only possible request would be to replace or fix munin, I think I broke it when I was messing with hostnames, it is sort of a delicate flower. However, your local pool routing idea is something I have toyed with doing myself, so feel free to work on that as well!

This message makes me happy.. the original goal of BAMT was to make deploying a new miner as fast and simple as possible.   Building, configuring and maintaining my own farm was a bit of a nightmare, and BAMT was created so that I would never have to do that again   I'm glad that it is accomplishing it's mission for you, as it has for me.

Thanks for the donation.  I was getting sort of burnt out on BAMT a while back, two weeks went by without a single donation (yet hundreds of downloads).  Sort of makes you question why bother to spend time on the project, when I could spend the same time making good money at work or just enjoying life.  But, in recent days several people have stepped up and provided either monetary support or volunteered their time helping with the project.   It's really not about the amount or what I make off donations compared to time spent... if I look at what I'm making per hour working on BAMT its a total joke.  It's just the concept of so many people taking and giving nothing back.  I can get a rough idea of the number of machines running BAMT by the number of hits on the fix distribution web site (no I am not tracking anyone   and if 1 in 100 has donated time or money, I'd be surprised.   Over the years I have worked on several OSS projects where there is absolutely no compensation, but creating software that is directly used to make money, and for no other purpose at all really doesn't feel the same to me, despite the source to BAMT being completely open.  Maybe I am wrong to expect people to give something in return for such software, who knows.

OK.. enough rambling, consider this a state of the project type of update heh.

Well since I've already gone on far to long, no harm in talking a little about the future of BAMT I suppose.

There are two major changes in store, one being the proxy based design mentioned earlier.  This will allow lots of nice things to be done with work management.  Miners can connect to a proxy instance on their own machine, or all your rigs can connect to a central instance, or any mishmash that works for you.  Failover, work distribution, pool changes are all done very nicely and without any restarts on the mining client.

The second major change is the web interface.  The current munin based stuff is fairly broken, and was intended more of an example of what you could do than as something to be used as is.
It doesn't even work a lot of the time.  I hadn't planned on doing much with that kind of thing at all to be honest.  I'm not a web guy.  However, Fitty has designed a really excellent web interface that makes configuring and tweaking your GPUs very easy.  Its impressive stuff, and I think unlike anything found anywhere else.  All credit goes to him.  This and the proxy should make their way into a "0.5" release in the next couple weeks, if all goes well.

ok.. that is a lot of crap for one message. i'm done.

[kirax]

...
I have nearly completed a very nice proxy that will run on BAMT, allowing you to do many advanced, interesting, and silly things with how your GPUs connect to your pools.   Basically the local GPUs just all connect to a local proxy instance.  The proxy takes care of making sure the work gets done and submitted where you want it to, and nobody sits idle.  The proxy can intelligently manage connection to pools without the miner clients ever needing to worry about it... they ask for work, they get work... that's all they need to know.

IMHO, the miner client should be a simple as possible.. it should just mine.   Fancy logic doesn't belong down at that level, there should be clear separation between the calculation and the work routing layers.   Phoenix fits into this type of design nicely.

This sounds great! And I concur with your overall strategy as it is elegant in its simplicity. Not to say it is simply done--I sure sure couldn't.

Your strategy is sound, and unix-like. Commands should do one simple thing, so I can route them through pipes, grep, and awk, to do fancy things. I appreciate your hard work on this project, I am currently deploying about 5 GH with BAMT, and the setup time is the least I have ever seen, so low I'm no longer bothering to image my drives and make backups, because I have to edit all of two files in about 60 seconds. As such, there is 2.5 BTC on it's way to you shortly (waiting for confirmations to my withdrawl to avoid fees). My only possible request would be to replace or fix munin, I think I broke it when I was messing with hostnames, it is sort of a delicate flower. However, your local pool routing idea is something I have toyed with doing myself, so feel free to work on that as well!

This message makes me happy.. the original goal of BAMT was to make deploying a new miner as fast and simple as possible.   Building, configuring and maintaining my own farm was a bit of a nightmare, and BAMT was created so that I would never have to do that again   I'm glad that it is accomplishing it's mission for you, as it has for me.

Thanks for the donation.  I was getting sort of burnt out on BAMT a while back, two weeks went by without a single donation (yet hundreds of downloads).  Sort of makes you question why bother to spend time on the project, when I could spend the same time making good money at work or just enjoying life.  But, in recent days several people have stepped up and provided either monetary support or volunteered their time helping with the project.   It's really not about the amount or what I make off donations compared to time spent... if I look at what I'm making per hour working on BAMT its a total joke.  It's just the concept of so many people taking and giving nothing back.  I can get a rough idea of the number of machines running BAMT by the number of hits on the fix distribution web site (no I am not tracking anyone   and if 1 in 100 has donated time or money, I'd be surprised.   Over the years I have worked on several OSS projects where there is absolutely no compensation, but creating software that is directly used to make money, and for no other purpose at all really doesn't feel the same to me, despite the source to BAMT being completely open.  Maybe I am wrong to expect people to give something in return for such software, who knows.

OK.. enough rambling, consider this a state of the project type of update heh.

Well since I've already gone on far to long, no harm in talking a little about the future of BAMT I suppose.

There are two major changes in store, one being the proxy based design mentioned earlier.  This will allow lots of nice things to be done with work management.  Miners can connect to a proxy instance on their own machine, or all your rigs can connect to a central instance, or any mishmash that works for you.  Failover, work distribution, pool changes are all done very nicely and without any restarts on the mining client.

The second major change is the web interface.  The current munin based stuff is fairly broken, and was intended more of an example of what you could do than as something to be used as is.
It doesn't even work a lot of the time.  I hadn't planned on doing much with that kind of thing at all to be honest.  I'm not a web guy.  However, Fitty has designed a really excellent web interface that makes configuring and tweaking your GPUs very easy.  Its impressive stuff, and I think unlike anything found anywhere else.  All credit goes to him.  This and the proxy should make their way into a "0.5" release in the next couple weeks, if all goes well.

ok.. that is a lot of crap for one message. i'm done.

On the note of additional help, I have some half-decent shell scripting experience, ancient programming experience, and five machines (gigabyte boards, AMD semprons, and 6870ref/non-ref, 5830, 5870 (asus non ref) mix) to test stuff on. Let me know if I can assist, I am an IT contractor but I have some spare time right... Well, after Tuesday or so, I hope.

[oo-oo]

Its someone else getting this error ?

user@miner:~$ sudo /etc/init.d/mine start
Starting mining processes...: mineYAML Error: Stream does not end with newline character
   Code: YAML_PARSE_ERR_NO_FINAL_NEWLINE
   Line: 0
   Document: 0
 at /usr/share/perl5/YAML.pm line 36