Bitcoin Forum

United Miners - http://www.UnitedMiners.com (http://www.UnitedMiners.com)

---

Managed rigs

•  To be used on a dedicated Linux rig you want to control with your phone/PC [BETA]
•  OC Vcore and memory with your finger from your phone
•  Automated temperature watch: miners pause if they overheat
•  Kick start your miner automatically with idle detection
•  Powerful statistics: stales and Mh/s reported to cell phone page
•  Remote reload/reconfigure/reboot from your phone or PC
•  Boot off USB thumb/SATA/IDE/SCSI , from 2GB up
•  Overheat email notifications, kick start email notifications
•  Settings configurable per rig like SMTP server, RPC L&P, fan speed %, over heat temp, OC core Mhz, OC memory Mhz
•     Managed Rigs HUD screenshots from an iPhone: http://imgur.com/BHl2a&1G5Js (http://imgur.com/BHl2a&1G5Js)  http://imgur.com/BHl2a&1G5Jsl (http://imgur.com/BHl2a&1G5Jsl)

• 0% fee, donations for extra services
• Offical DNS Seed for the bitcoin client being introduced in 0.4 (dnsseed.unitedminers.com)
• iPhone/Android web site access (with API Key) to your managed rigs
• SMS/txt notifications on idle miners
• Completely redundant backend systems, we could lose server motherboards and you would never notice
• VM's float on host nodes in the cluster without losing a packet
• Active/Passive and Active/Active expandable clusters for all services
• Passwords are secure. Hashed, sh1, then salted.
• Classic worker per rig system fully supported
• Long-polling support
• JSON API access for user stats, worker stats, and pool stats
• Top lists and pool stats
• Personal account Mh/s mining graphs going back up to a year
• Proportional payout method
• On-demand payouts.  (After a block has been confirmed)
• 8 decimal place payouts
• Instant pay out button (0.01 BTC min)
• US West coast based data center
• IRC support at #UnitedMiners on freenode or from the website with our web IRC page
• Proportional payout method based on shares submitted per round
• SSL setup on the web server

We have been up and running for a while privately and are opening the doors to the public.[/list][/list]

Wow. Finally some unique and useful features in a startup pool. That remote control scheme is definitely a world-first.

The only thing that stands out negatively is prop. payments.
As you might know it's one of the worst models for a 24/7 non-hopping miner, any plans of PPS variant schemes or PPLNS?

Yes, already working on that on the dev servers. I don't want to reinvent the wheel so is PPLNS better/fairer then PPS?

Obv. pure PPS is ideal for the miner since he risks nothing and gets zero variance day in day out..
But it's not viable for you the operator unless you have massive reserves of BTC in store for a stormy period. Even then you'd need some sort of fee to fence off your risk, deepbit for example has 10% and rfcpool.com has 7%

PPLNS doesn't require any buffer on your part. It also can't be exploited statistically by hopping in after many high percentile rounds to profit from early, valuable shares of likely upcoming short rounds.

Short reading
http://eligius.st/wiki/index.php/Pay_Per_Last_N_Shares

* Passwords saved with AES256

AES256? Specs on that? So you'll able to decrypt and view users passwords?

Sorry for the confusion, I need to update the page. Passwords are hashed, sha1, then salted.

Before the site gets big (which i hope it does, nice features btw!) - switch to SHA256 or SHA512.
There is no reason for using SHA1 anymore.

A bit more database usage, but it's not even noticeable with 1m users (the difference between sha1 and sha512 is 88 bytes per user, 88 000 000 bytes for 1 million users,  thats 83.9 MB difference.)

Thanks!

No worries, may i ask what you're using for load balancing / failover?

SSL enabled on the web site now.

Curious is your iphone / android dealio an actual app or is it a webpage frontend?

If its an app, have any plan to make a webos app? =)

Web page designed for an iPhone with big buttons and that "look". Native app would have been nice but then it wouldn't work on anything but iOS. The Managed Rig page in the screen shots should work find on webOS.

You say "Proportional payout method based on shares submitted per round". Now when I read the website FAQ it says

"How are payouts calculated? [TOP]
We use the Pay Per Last N Shares method to calculate payouts when a round ends. "

You should not change the payout method behind the back of people, and not during a running round. Also N is not specified!

It is kinda messed up changing payout schemes in the middle of a round, invalidating work people have already put in

How are you going to pay the miners that submitted proportional shares (like me). You can't change the payout method in the middle of the round. I had a positive balance and now it's gone. Please do the right thing and take care of this.

After talking to other pool Op's I've decided to move to PPLNs after this round. It will go into effect when the round ends and I've already updated the FAQ. The payout code hasn't been change but the payment estimate code has been when we moved over to 0% fee with donation levels.

If you want to discuss sticking with straight prop lets talk about it, I'm open. The reliable 24x7 miners appreciate PPLNS because it can't be gamed so easy like proportional can be.

Straight prop.
I'd be willing to donate premium percentage, so say like 2.5% to 5% to have hopping access along with a reliable API.  Tie the pool stats into a the user API, which can only be enabled after you donate a predetermined percentage.  You can then redistribute this fund to your full time miners as "additional hopper income".  I think the ~50GH/s will definitely help a pool out and also attract more miners.  Even if most hoppers leave after 43% of shares being submitted, there is still an incentive for finishing a block, because without the block being finished no one will get payed hopper or not.
http://www.digbtc.net/ DigBitcoin resurrected itself after switching back to purely prop after trying LNSMPPS.

I like this idea. Has this worked out for any other pools? Do hoppers think it's a good idea?

I'm not sure, I'll ask though.  The closest implementation is by bitcoinpool.com.  If you donate any amount they will not flag you as a hopper.  If you get flagged as a hopper you will lose 50% of your earnings and they will be redistributed amongst the pool. 

Yes we do. I donate 1% where ever I mine if they have it set up. Many of us have thought about offering donations to ease the anger some.

Some pools also do something different

They offer both prop and PPS.. (dont know if you can handle PPS yet)

anyways when they do that, we can use their PPS as a back up pool.. which we jump on when no one else is lucky.. which is pretty often sometimes.

The hopper is coded so you can set it up to jump on the backup pool with the most shares, when you are on backup.

This could also help you in a way, as hoppers get the bad rep for only being around on short rounds and being very missed on the very long rounds.

Sometimes we even help pools that dont have PPS set up.. a bunch of us jumped on ozco.in when they had a 9 million share  block and helped

 get over the top at 10 million and found the block. It could have been a few more days without all the extra hashes.


just a thought anyways. We like people who dont hate us.

I'd be up for that.   Happy to pay a donation to the site as mentioned above.  Everyone's a winner.

Donate. I always do. 1% no prob. More if reliable like fast api.

I also like your pool from the looks, would be a shame if you did not get rewarded for it.

1) I am reading your FAQ page right now and:
"How are payouts calculated? [TOP]
We use the regular proportional method. We are testing and considering a Pay Per Last N Shares method to calculate payouts when a round ends if the account is flagged as a hopper. We are also considering hopper flag protection at a certain donation rate. Feedback is greatly appreciated! "

So, what are you using right now? Regular proportional or PPLN?

2) Oh, and BTW, there is no need to hop away at >=43% if you have no pools at lower % you currently are to switch to (can a grammar nazi fix taht for me) :) .  Organofcorti made some cool simulations and calculations to prove it: https://bitcointalk.org/index.php?topic=26866.msg439864#msg439864

My balance said .08xxxxx and then I did the instant payout and all of a sudden my wallet has .078 more.  Hmmmm.

Whats your user ID number?

EDIT: (was this you?) ...
        "category" : "send",
        "amount" : -0.07863839,
        "fee" : 0.00000000,
        "confirmations" : 7,
        "txid" : "7b5db8ccea52b38d1e0e992c7e4bda1881e7839248ea9d67a7318b3fb21d2ea2",
        "time" : 1312933330

You balance was 0.07863839 with a 0% donation.

be having a lot of connection issues.
I have a 2.5% donation set for QOS
why do I still have connection issues?

What are you using for OC's and fan speed?  Is it not possible to set per GPU? And does anyone know how Diablo compares to phoenix?

Right now the userland programs are setup to be able to OC all GPU's with 1 speed or each GPU at a separate speed but the WWW GUI is setup for 1 speed for all. Since there isn't much interest in the managed rigs I don't know if I want to spend the time adding that feature.

The userland programs on the linux image are also setup for either Diablo to manage all the GPU's or 1 phoenix miner for each. However the WWW GUI just uses Diablo. I'm still the only user with managed rigs so it's probably going to stay that way right now. Being a pool op of a 0% fee pool and competing with BTCGuild and deepbit is a thankless job.

Well I hope it improves, it's a great idea.  I prefer to stick with PPLNS pools and I've been using Linuxcoin and BAMT but the iPhone and web management would make things much easier.  I'll be happy to throw a couple Ghash your way at 3.5% to test it out as soon as I sort out a small lightning issue - didn't think to run cat5 through surge protector.

why are there so much stales?
(7.26%)
TIA

We just switched to luke-jr's version of pushpool this afternoon, hopefully that clears some up. Bitcoind is always an issue also and when it has problems and needs to be reset you see some stales/rejects.

UPDATE - wont be using luke-jr's pushpoold from git. The stale/reject rates where crazy! Switched back to the one I know works well. My test rig still has 0% stales so that's a lot better. The new LB makes life a lot easier to.

so, now I stop
here I see:
UnitedMiners
shares 4261    stales 1806   (42.36%)!!!
and on your HP I see
2430 shares.... >:(

My rig I see 3337 accepted 98 stales, or about 2.9%. Thats not so bad now. Reset your counters on your miner. The luke-jr pushpoold test was a disaster.

Anyone else seeing a lot of upstream RPC errors for workers and problems connecting to the API?

 ???