Raspberry pi (MinePeon) & 9 eruptors, can I add a cube to it?

[J_Dubbs]

I've got this little guy on my desk:


But across the room on the same network I have cubes and blades. Most of the cubes/blades are pointing at proxy running on desktop PC, and I've got a cube (via crossover cable) and BFL Jally running on an old laptop, kinda can't push the WiFi on the laptop much more than that I don't think.

Anyways, is it possible for me to run a proxy on MinePeon (BFGminer) and point a cube at it while still running the erupters? or will that overload the little WiFi in the Raspberry pi?

I'm trying to weigh my options to deal with some potential network traffic clutter...

Thanks!

[J_Dubbs]

Just a follow up, yes this works, and in fact it improved the efficiency on my cube.

Added the following switches under Minepeon BFGminer settings: --http-port 8337 --net-delay

Then I pointed cube at raspberry IP addy with port 8337 specified.

Not sure what the net delay one does, but it seems to be working. At first one of my erupters got knocked off but I eventually got it working thought eventually rebooting.

[kfactor]

I'm running blades and cube on a pi right now without --net-delay and I have no issues.

Well, one issue... I am desperately waiting for bfgminer to support antminer u1's so I can throw them on the pi too (and hopefully it doesn't choke)

[J_Dubbs]

I'm running blades and cube on a pi right now without --net-delay and I have no issues.

Well, one issue... I am desperately waiting for bfgminer to support antminer u1's so I can throw them on the pi too (and hopefully it doesn't choke)

I had to remove my cube from it because the RPi was dropping connection between rounds. I removed the --net-delay line, not sure if that is what messed it up(?).

[drmadison]

I currently have 120GH/s worth of BFL gear, 2 cubes, and 4 Antminer U1s going through 1 RPi so far...no issues I've seen (running across 2 cgminer and 1 bfgminer instances).

[J_Dubbs]

I currently have 120GH/s worth of BFL gear, 2 cubes, and 4 Antminer U1s going through 1 RPi so far...no issues I've seen (running across 2 cgminer and 1 bfgminer instances).

Are you using MinePeon?

[drmadison]

No, I'm just using Raspian

Compiled cgminer initially (following a guide I found...somewhere via google) and used that for the BFL gear.
Later added some Cubes and for whatever reason slush's stratum proxy (stratum_proxy.py) wouldn't connect for me, so I compiled a copy of bfgminer to use as the proxy.

Then I recently got a small number of Antminer U1s and they needed a new copy of cgminer - since it wasn't part of the main github I went ahead and compiled it separately (with JUST Antminer support) so I wouldn't risk breaking anything else.

[J_Dubbs]

No, I'm just using Raspian

Compiled cgminer initially (following a guide I found...somewhere via google) and used that for the BFL gear.
Later added some Cubes and for whatever reason slush's stratum proxy (stratum_proxy.py) wouldn't connect for me, so I compiled a copy of bfgminer to use as the proxy.

Then I recently got a small number of Antminer U1s and they needed a new copy of cgminer - since it wasn't part of the main github I went ahead and compiled it separately (with JUST Antminer support) so I wouldn't risk breaking anything else.

Man, I wish I knew why mine was falling offline between rounds. Aside from that it was working awesome.

[kfactor]

try changing the port on cube, blades and bfgminer to 8330 - I know it sounds weird but try it and see what happens ...

[J_Dubbs]

try changing the port on cube, blades and bfgminer to 8330 - I know it sounds weird but try it and see what happens ...

I'm actually going in a different direction, but I have another cube showing up later this week and will revisit this at that time. I was running it on 8337, fwiw, not sure if that matters... Almost certain I hadn't tried 8330.

[J_Dubbs]

try changing the port on cube, blades and bfgminer to 8330 - I know it sounds weird but try it and see what happens ...

Okay, got my new cube and it seemed to be slowing down too much on the laptop with the other 2 Slush Proxies and BFG miner, maybe too much network traffic for the laptop to handle?

Anyways, using the Raspberry with Minepeon, added the switch to look for 8330. It's working at a good speed, still seeing my other blade a bit slow but it might need more time. Now I just need to wait for a round to end so I can see if it drops connection. Will post updates as they unfold.

[J_Dubbs]

Ok... so I changed some things in my whole config. Disabled the WiFi on the laptop and removed the bridged connection. Instead of hooking one cube up directly to the laptop I'm running them all through my switch, laptop is also cat5e to the switch now too. My cubes running to the laptop are slowly climbing, not sure if they will pass 30gh/s but the efficiency stats look like they eventually will. The cube pointed at Rpi w/Minepeon fired up to full hash very quickly, but I'm not sure if that skews the difficulty for my USB eruptors that are connected to it.

I have another Raspberry Pi on order, when it arrives I plan on running Minepeon and hopefully getting all 3 cubes to run on it. I know that Slush's proxy does something weird with multiple cubes on a single instance, and thus the efficiency gets reduced, and I also know that BFGminer doesn't handle the cubes best while running as a proxy. I'm just not sure if Minepeon will have any issued if I point 3 cubes at it, any insights on this?

EDIT: also seeing despite the fast warmup there's kinda a lot of rejects on the cube pointing at Raspberry Pi/Minepeon. "Jobs:0000021752 Accepted:0000020269 Rejected:0000000935"- dunno, does that seem high? Efficiency is 93.18%, hash is 32,014

EDIT2: Moving the 3rd cube back to the laptop to see if it does better there now that everything is off WiFi and running through the switch...

EDIT3: Yeah, seems like I get a lot less rejects by running through the laptop and a Slush Proxy with a unique port. Everything has good efficiency, low rejects, and decent hash speeds now.

[razorfishsl]

I currently have 120GH/s worth of BFL gear, 2 cubes, and 4 Antminer U1s going through 1 RPi so far...no issues I've seen (running across 2 cgminer and 1 bfgminer instances).

just a tip…..

Run it for 8 hours on the Pi, then plug it into a real computer for 8 hours and report back your findings,

I've see instances of 'A' shared being significantly down, even though the speeds of the miners match.

[drmadison]

My pool is reporting as-good-as-expected (occasionally better than expected) hash rate so I think it's all good (and CPU usage is < 50%).

I have 2 50GH/s units (running at 57GH/s each), 1 7GH/s unit, 1 Cube and 3 Antminer U1s hooked up right now, and I'm pulling (according to my pool) 156 GH/s
57+57+7+30+4.8 = 154.8 so I'd say I'm in spec definitely...

And before swapping to the RPi for all of this it was going through a Surface Pro - but it constantly rebooted so that got scrapped. Overall no noticeable difference though!

[J_Dubbs]

I currently have 120GH/s worth of BFL gear, 2 cubes, and 4 Antminer U1s going through 1 RPi so far...no issues I've seen (running across 2 cgminer and 1 bfgminer instances).

just a tip…..

Run it for 8 hours on the Pi, then plug it into a real computer for 8 hours and report back your findings,

I've see instances of 'A' shared being significantly down, even though the speeds of the miners match.

This might line up with what I was seeing. I mean, the cube on the Raspberry had full hash and got up to speed fast, but the number of rejects reported in the cube's web UI were way higher than when running through a PC. Eh, if I get another Raspberry I might just play games on it instead for a while, since it's in the mail already I guess that's what happens next...