A Dell PowerEdge T430 is reporting several voltage-related alerts through the management console, including system board fail-safe voltage, CPU 1 VTT and VDDQ power-good voltage, and CPU 2 VCORE power-good voltage being outside the expected range. The problem usually appears after the client loses power. The server then may refuse to start until the power supplies are disconnected and the system is fully drained. On a few occasions, removing and reseating DIMMs seemed necessary before it would boot again. BIOS and management firmware are current. Could this be caused by the RAM, power supplies, configuration, or a failing system board?
3 Answers
This sounds primarily like a system-board issue, although a marginal PSU or power-distribution component could produce similar symptoms. You can strip the server down and add components back one at a time, but since the RAM and PCI cards have already been reseated without solving it, I’d prioritize checking the PSUs, power-input events, and management logs. If a PSU swap does not change anything, I’d plan on replacing the board.
The power-loss pattern and the fact that draining the system brings it back are the biggest clues. Multiple VTT, VDDQ, VCORE, and board-level power-good faults across both CPUs point more toward power sequencing or voltage-regulator trouble than bad RAM. The DIMM reseat may be coincidental; removing power is likely what actually clears the fault. Start by testing with a confirmed-good PSU, then test each installed PSU separately. If the errors continue, the system board or its VRM circuitry is probably failing.
That sounds like the most practical test I can do first. I’ll try known-good power supplies and compare the results.
A failing management controller is possible, but the voltage power-good alerts are more consistent with a PSU, power-distribution board, or VRM problem. Check the management logs for PSU and input-power events around the failure. Since the server is out of warranty, it would be worth testing known-good supplies before buying a replacement system board; older servers have been known to require both a board and the PSU cage or power-distribution assembly.
I’ll check it with confirmed-working power supplies first. Other T430 systems here have not shown the issue, so this appears isolated to this machine.
I’ve already tried removing and reseating the memory and PCI cards, and the problem keeps returning. I haven’t removed every cable, but the board or power delivery seems more likely now.