All miners dropped to 50% last night

[gnaget]

I normally run at around 7 GH/s when everything is running perfectly, but I woke up this morning to only 3 GH/s.  Looking at my workers (split up by machine), they were all at about 50%.  I should have checked the individual miners to see which were locked up, but running that slow put me in a panic.  A quick reboot on all the machines and I was back up to 7.

I'm used to having miners lock up on a single or maybe 2 machines, but never all of them.  I would think maybe the heat got real bad, except one of my miners is completely separated from the others, and it was not even immune.  All of them are running Ubuntu with phoenix miner 1.5. 

Have any of you run into a similar problem?  What did you do to fix it?

[Arius]

It's possible that the problem was in your fuse box.  I run several dozen computers and I've noticed that if a 30 amp fuse starts to overheat, the heat moves to the neighboring fuse and there are power anomalies in my house, clocks keeping bad time, etc.  If all your miners are cranking out at peak performance, your house may not be able to deal with the heat in the internal wiring.

[AngelusWebDesign]

7GH/s is a *lot* of mining capacity to be running in a residence, unless you don't have to run air conditioning (which is true for very few Americans in July).

[gnaget]

7GH/s is a *lot* of mining capacity to be running in a residence, unless you don't have to run air conditioning (which is true for very few Americans in July).

Residence, apartment, same thing...  I know it is a lot, and I am definitely at a wall with how much I can handle.  Any more, and I have to hire an electrician which I already got permission to do.  Did I mention I am located in Texas and we have had seemingly non-stop 100+ days for the last 2 months?

Luckily, my apartment is very modern, and I have the power distributed across 5 different circuits with 15 amp extension cables.  I have them distributed well, so I don't trip breakers anymore, so I don't think I am overtaxing them.

[NetTecture]

7GH/s is a *lot* of mining capacity to be running in a residence, unless you don't have to run air conditioning (which is true for very few Americans in July).

Residence, apartment, same thing...  I know it is a lot, and I am definitely at a wall with how much I can handle.  Any more, and I have to hire an electrician which I already got permission to do.  Did I mention I am located in Texas and we have had seemingly non-stop 100+ days for the last 2 months?

Luckily, my apartment is very modern, and I have the power distributed across 5 different circuits with 15 amp extension cables.  I have them distributed well, so I don't trip breakers anymore, so I don't think I am overtaxing them.

Americans

In germany every new build housing unit can - per spec - get 63 amperes on 3 phases, 230 volt. A nice 63kw Normally a housing unit starts with 35 ampere phases in most areas

[Kermee]

Americans

In germany every new build housing unit can - per spec - get 63 amperes on 3 phases, 230 volt. A nice 63kw Normally a housing unit starts with 35 ampere phases in most areas

Most newer homes in my area have service inputs of 200A, two-wire 3-phase 240V/60Hz, including my home.

The main problem for 'Americans' is that almost everything operates at 120V meaning we're using twice the amperes to run electronics to achieve the same power use at 240V.

Actual serviceable power isn't the issue for most Americans.

I have 18 rigs running on six 120V/20A circuits, which if I fed the PSU's 240V instead of 120V, I could of run 18 rigs on only two 240V/30A circuits... 120V is retarded but we're stuck with it in the U.S.

Cheers,
Kermee

[Xephan]

It's possible that the problem was in your fuse box.  I run several dozen computers and I've noticed that if a 30 amp fuse starts to overheat, the heat moves to the neighboring fuse and there are power anomalies in my house, clocks keeping bad time, etc.  If all your miners are cranking out at peak performance, your house may not be able to deal with the heat in the internal wiring.

I don't think an "overheating" fuse has anything to do with this at all. By design, if the current exceeds what a fuse is designed to allow, the circuit should trip. There's no in between where it somehow can tell your computer to run at half speed.

[Xephan]

I normally run at around 7 GH/s when everything is running perfectly, but I woke up this morning to only 3 GH/s.  Looking at my workers (split up by machine), they were all at about 50%.  I should have checked the individual miners to see which were locked up, but running that slow put me in a panic.  A quick reboot on all the machines and I was back up to 7.

I'm used to having miners lock up on a single or maybe 2 machines, but never all of them.  I would think maybe the heat got real bad, except one of my miners is completely separated from the others, and it was not even immune.  All of them are running Ubuntu with phoenix miner 1.5. 

Have any of you run into a similar problem?  What did you do to fix it?

Did you split each GPU between the same X number of workers e.g.
GPU A -> Worker A on Pool A
GPU A -> Worker B on Pool A
GPU B -> Worker A on Pool A
GPU B -> Worker B on Pool A

If so, then if one of the GPU locks up due to heat, you will see half the throughput across all worker accounts. Which was what happened to me before. Increasing fan speed fixed that.

[gnaget]

Did you split each GPU between the same X number of workers e.g.
GPU A -> Worker A on Pool A
GPU A -> Worker B on Pool A
GPU B -> Worker A on Pool A
GPU B -> Worker B on Pool A

If so, then if one of the GPU locks up due to heat, you will see half the throughput across all worker accounts. Which was what happened to me before. Increasing fan speed fixed that.

I have it distributed (only one pool):
CPU A, GPU A -> Worker A
CPU A, GPU B -> Worker A

CPU B, GPU A -> Worker B
CPU B, GPU B -> Worker B

etc.

Everything is running smoothly now, I can only assume it was a really random event, but given that they all dropped in speed within the space of an hour, it was scary. 

[m0w3r]

I had a 50% drop for all cards on one rig before.  The other fig (on same circuit) was fine.  Also, all was fine on reboot.

[Arius]

That's funny, I'm in Texas too.  It has been hotter than hell the last few days.  I'm figuring the next CPS bill will be around $900.