What usually fails in your RIG after 6+ months??

mikeb99

My 5xRX480 Ethos gpushack rig has been running great for last 6 months. However, just started getting random errors, with some GPUs dropping hashes even hanging/crashing after 1-2 days of reboot.

What are the components I should be looking at / inspecting to replace? Risers maybe? Would like to get a sense of what components are usually the first to go.

Thanks!

Procrastitator

I've only had power supplies go bad in my long term rigs.

suthek

fans can go.
In the summer I had cards degrade and clocks that previously worked no longer work.
I believe it was due to running them too hot and causing solder traces to melt.

What temps are you running the cards at?

headshot155

I had similar issues on one of my rigs, dropping hashes etc, getting erro 43 a lot. Thought is was risers but in the end all it took was a fresh windows install. It’s probably the easiest thing to try first

Jukebox

Risers due to bad +12 V power connection.

lablett

I would run hardware info check the number of hardware errors. How dusty are your cards now as that can have massive impact if your GPU fins are now blocked. Weak components will usually fail in the first couple of months. I would like to know what model your power supply is and how much it is pulling at the wall ? As a quick check can you go into GPUZ and tell us how much power each card is pulling. Also can you watch the levels to ensure they are fairly constant.

glenn_tx

Primarily burned out riser power cables (molex, etc) Literally melted. I made the mistake of buying a few Rosewill 1000 watt PSUs. GPU power cables burned out at the PSU. If a GPU stops working, 1st thing I try is swapping out the riser and corresponding cables. 1/2 the time, that fixes it. As someone else said, reinstalling Windows is always a good thing to try. I can rebuild a machine in a few minutes with my eyes closed.

I've had to two GPUs die completely. That was due to putting an Arctic Cooler after-market cooler on my GPU. Coolant ended up leaking and ruining my card. Grrrrr.

mikeb99

Thanks everyone! I'm running EthOS, what are my options to check/scan for hardware errors?

Also I am thinking its the risers, may just replace them all.

BUTUZ

6-12 months? PSU fan or GPU fan. Nothing else should fail. Only moving parts. If your risers, with no moving parts fail in 6 months then you bought crapola ones.