F5F Stay Refreshed Hardware Desktop The server power supplies are malfunctioning repeatedly.

The server power supplies are malfunctioning repeatedly.

The server power supplies are malfunctioning repeatedly.

W
Willz_03
Member
179
12-10-2022, 02:04 PM
#1
My organization operates several servers equipped with multiple GPUs. Two of these units failed within a week. The hardware details are listed below. On Monday, one server stopped working after replacing its power supply and connecting the motherboard and CPU. It booted normally, then failed to start again after being turned off. After removing the GPUs, it still wouldn’t power on. Later, when only the new power supply was connected, the second server also stopped working. Its specifications match the first unit: ASUS Z690 Prime motherboard, i7-13700K CPU, Corsair Vengeance LPX 128GB RAM, Noctua cooler, MSI Gaming Trio 4090 GPU via riser cable, Silverstone Hela 2050 Platinum PSU, and Ubuntu 22.04 OS. Any assistance would be greatly appreciated.
W
Willz_03
12-10-2022, 02:04 PM #1

My organization operates several servers equipped with multiple GPUs. Two of these units failed within a week. The hardware details are listed below. On Monday, one server stopped working after replacing its power supply and connecting the motherboard and CPU. It booted normally, then failed to start again after being turned off. After removing the GPUs, it still wouldn’t power on. Later, when only the new power supply was connected, the second server also stopped working. Its specifications match the first unit: ASUS Z690 Prime motherboard, i7-13700K CPU, Corsair Vengeance LPX 128GB RAM, Noctua cooler, MSI Gaming Trio 4090 GPU via riser cable, Silverstone Hela 2050 Platinum PSU, and Ubuntu 22.04 OS. Any assistance would be greatly appreciated.

T
TheKillerMomo
Junior Member
48
12-10-2022, 02:46 PM
#2
The wall outlet can handle 2000 watts, but using a single power supply isn’t ideal. You’re likely aiming to transport all your groceries in one trip, not just for convenience.
T
TheKillerMomo
12-10-2022, 02:46 PM #2

The wall outlet can handle 2000 watts, but using a single power supply isn’t ideal. You’re likely aiming to transport all your groceries in one trip, not just for convenience.

T
Thifleno
Member
206
12-13-2022, 02:25 PM
#3
In DC yes, there was no problem before now and it won’t happen again. Both of them initially failed at the DC, and then a replacement died when it was returned to the office.
T
Thifleno
12-13-2022, 02:25 PM #3

In DC yes, there was no problem before now and it won’t happen again. Both of them initially failed at the DC, and then a replacement died when it was returned to the office.

N
nathanmizzi
Member
127
12-13-2022, 05:16 PM
#4
it appears the variable can represent location while the constant defines where and how it's connected to both walls and components. examine the devices and check the power interfaces
N
nathanmizzi
12-13-2022, 05:16 PM #4

it appears the variable can represent location while the constant defines where and how it's connected to both walls and components. examine the devices and check the power interfaces

C
Chester709
Junior Member
44
12-13-2022, 09:42 PM
#5
I’m still trying to figure out why my power systems are failing. Both servers shut down at the same time, but one stopped right after I swapped the PSU at the repair shop. Those units are supposed to last, not just break down quickly. I’m leaning toward checking what caused the PSU failure on the first server, since that’s where the real clues might be.
C
Chester709
12-13-2022, 09:42 PM #5

I’m still trying to figure out why my power systems are failing. Both servers shut down at the same time, but one stopped right after I swapped the PSU at the repair shop. Those units are supposed to last, not just break down quickly. I’m leaning toward checking what caused the PSU failure on the first server, since that’s where the real clues might be.

J
JamyDodger2
Junior Member
30
12-14-2022, 05:24 AM
#6
Confirm the replacement: swapping the PSU usually means swapping all the cables too. Are you certain there aren't any faulty cables elsewhere?
J
JamyDodger2
12-14-2022, 05:24 AM #6

Confirm the replacement: swapping the PSU usually means swapping all the cables too. Are you certain there aren't any faulty cables elsewhere?

X
XxAlenxX
Member
118
12-14-2022, 10:53 AM
#7
OP works in a data center, not a personal computer builder. You should request them to purchase suitable testing gear. PCIE lane testers, power analyzers, thermal cameras are helpful tools that could have avoided this issue altogether.
X
XxAlenxX
12-14-2022, 10:53 AM #7

OP works in a data center, not a personal computer builder. You should request them to purchase suitable testing gear. PCIE lane testers, power analyzers, thermal cameras are helpful tools that could have avoided this issue altogether.

H
HorseGirl9245
Junior Member
41
12-17-2022, 01:48 PM
#8
I’ve reviewed all the cables for faults, yet I haven’t swapped them out—just updated the power source. Looking back is always clear. My role isn’t in a data center; I’m a sys admin in an office with servers in a local data center. I’m also uncertain about the tools required, as this isn’t something I’ve tackled before and haven’t dealt with electrical problems on GPUs.
H
HorseGirl9245
12-17-2022, 01:48 PM #8

I’ve reviewed all the cables for faults, yet I haven’t swapped them out—just updated the power source. Looking back is always clear. My role isn’t in a data center; I’m a sys admin in an office with servers in a local data center. I’m also uncertain about the tools required, as this isn’t something I’ve tackled before and haven’t dealt with electrical problems on GPUs.