How do I troubleshoot intermittent fan surging on UCS22

[RedNectar]

Friends,

I'm going to need some help soon. I have a customer who has a few UCS C22 - a little bit old, but here's the deal.

Following a power surge during an electrical storm, one of the servers is going through a cycle of full-speed fan/reduced speed fan.

We've established that when the fan speed increases, power consumption surges (that sounds logical).  My thinking so far is:

• repeat {
  • something is overheating
  • this causes the fans to go to full speed
    • this causes the power surge
  • the overheated component cools a little
  • fans slow down
• }

So far he has tried swapping power supplies and fans with a spare unit but the cycle continues. And since the underlying OS is Windows, he's tried rebooting the box a few times too.

So my question is, how do I figure out what is overheating? Or could it be some other problem causing the problem?

I hope someone can help.

---

---

[Kirk J]

Greetings.

SSH to the CIMC

connect debug-shell

#sensors

Look  at the various thermal sensor readings, threshholds, and see if anything stands out.

Stand alone C series, especially on older code are not very helpful with tracing down thermal/fan events.

If you have a raid card, if the internal ROC chip temp hits 85 C, it also revs the fans pretty high.

Is this server sitting along side other servers not hitting the issue?

Assuming we haven't actually hit some sort of environmental/HVAC issues?

Kirk...

[RedNectar]

Thanks Kirk,

I'll give that a shot

Chris