My RIG is Restarting every 15 minutes

[josh75]

Hello all,

i have a RIG of 12 GTX 1070's

i am actually mining ETH.

unfortunately it is restarting every 15 minutes wich make me loose at least 1 minutes of mining.
so i loose approximatly 90 minutes of mining per day.

i use simple mining and ubuntu.

is someone know why is this happening thank you all

[1715390405]

1715390405

[1715390405]

1715390405

[joshuajones02]

might be better off asking in the simple mining thread.. could be a number of different possibilities. If its a hardware issue, it'll be hard figuring it out with that big of a rig

[adaseb]

Most likely one of your risers is messed up.

[minerlai]

Unplug your GPU 1 by 1 until you achieve stability then start adding back.

What I suspect is

1) Too much power draw
2) faulty riser/GPU

You need to isolate by testing.

[Mrvain]

Hey Josh.

I have the same problem here with my 12 GPU 1070 (Asrock H110 pro Mobo) rig using Simplemining on Ubuntu. The rig will run fine for anywhere between 5 - 60 minutes, and then crashes. Sometimes it restarts. Sometimes it crashes and freezes. My cards are not overclocked (watts limited to 125) and I have ample power with dual 1600w EVGA Platinum power supplies. I would like to know if you get this figured out.

[wacko]

I have the same problem here with my 12 GPU 1070 (Asrock H110 pro Mobo) rig using Simplemining on Ubuntu. The rig will run fine for anywhere between 5 - 60 minutes, and then crashes. Sometimes it restarts. Sometimes it crashes and freezes. My cards are not overclocked (watts limited to 125) and I have ample power with dual 1600w EVGA Platinum power supplies. I would like to know if you get this figured out.

There's plenty of reasons this could be happening. Power is less likely since you've got two high quality PSUs (but still one of them could be defective). Could be the board/cpu/ram issue (need to test them before building an actual rig, memtest for ram, typical cpu stress tests for cpu/ram like prime95 etc). One (or more) of the pci-e slots on the MB might be defective. If the platform (mb+cpu+ram) is stable and it's not the power issue then it's either cards or risers. Risers are way more likely since if there was something with your cards then the system/miner would crash sooner. But then again, even a single unstable card can bring the whole system down. Running a 12 gpu rig is tricky, too many components, and if there's something wrong with just one of them — the whole thing becomes unstable. The solution is to test all of them one by one. Either by removing things from the current configuration, or starting from just one gpu and then adding cards/risers by one.

[josh75]

Hello all,

i did put more than two riser per sata cables that is why i think.

i have change this to only 2 per sata or molex

after 24 hours it as only restart one time.

do you think i still have a problem

my rig is doing 1 restart every 24 hours approximatly is it bad ?

thank you all for your help and please tell me if your rig do this too.

thank you

what is also strange is i still have 2 gtx 1070 who run at 26 mh when the others are running 30.5mh with the oc

[wacko]

i did put more than two riser per sata cables that is why i think.

i have change this to only 2 per sata or molex

after 24 hours it as only restart one time.

It's usually not recommended to connect more than 1 riser with one molex/sata cable. Whether or not one cable will be able to handle more than 1 riser will depend on the quality of the cables/connectors, and the type of cards (some draw very little power from the pci-e slot, while others draw a lot, sometimes even more than they're supposed to, like it was the case with the reference rx 480 cards). You never mentioned the kind of PSU(s) you're using.

[josh75]

i have 1 HX 1200 Corsair
and 2 RM 1000 Corsair

for a 12 GPU rig wich is a lot of PSU

i have see one this link https://docs.google.com/document/d/1SvtkI_ZCHAxqbIsaWnuA81hRbI3Tp2Zf3iyEQlORLTU/edit

Why does SMOS Crash?
Don’t use low quality USB stocks. Some mining software uses more disk access than others (Claymore uses RAM where Etherminer uses disk) and will wear them out quick. The price of a good USB is the cost of a 120GB SSD, I would go with the SSD.
I use SSDs only the cost just a little more than USB, note in Etcher need to click settings then “Unsafe mode” to show SSD in the list of devices
Don’t use anything less than a GOLD level PSU or put more than 80% MAX load unless you will not have a stable RIG. Make sure to check power using an inline watt meter.
DO NOT put more than two risers per Molex or SATA cables it will get to hot and cause problems or physical damage


si i have change to 2 per sata or molex.

you think i should switch to 1 per molex or sata ?

what is strange is that my other rig as no restart problem with 2 sata or molex

but this one restart  one time every day randomly.

thank you all for your help

[josh75]

Does your rig restart a lot ?

do you guys think i should try one GPU per SATA or molex ?

Best regards

[wacko]

Does your rig restart a lot ?

My rigs almost never restart for no reason, they usually run for weeks and I'm the one that's restarting them (for driver or OS updates etc). Sometimes a rig starts misbehaving, and then I just figure out why it does that and fix the problem. Usually it all boils down to is just changing a faulty riser.

do you guys think i should try one GPU per SATA or molex ?

If you have enough cables then just go for it and see whether it helps. Though if you have another identical rig and that one works fine then it's probably something else. I outlined the basics of identifying the problem in my first post in this thread.