Replacing PSU shutdown the UCS

[peter.matuska1]

Hi,

customer has few UCS C240 M5 at 2 different location (new one and the old one). The servers were bought at different time. Every server has 2 PSUs. The steps taken are as follows:

1. remove the power cord from the PSU - server is UP, no error light on the front panel

2. remove the PSU from the server - server is UP, no error light on the front panel

3. insert the PSU to the server - server immediately shuts down

We have the same UCS in our lab with 1600W AC PSU and I couldn't replicate the issue. I upgraded ours to the same CIMC level and no issue.

This was tested at the new and the old location. The PSU is 1050W AC. One PSU is new (RMA) and the behaviour is the same. The test was done when the UCS was powered from the UPS and without the UPS.

Has anybody have this issue before?

thank you

[Kirk J]

Definitely need to open a TAC case on that.

I would want to take a look at the logs.

I am assuming the 2 PSUs are exactly the same PID?

Both AC sources are the same (i.e.110 vs 220 )?

Any GPUs installed?

What HUU/firmware version?

Thanks,

Kirk...

[peter.matuska1]

Hi,

yes, all PSUs are the same - Cisco PN: 341-0638-02-A0. The newest one (RMA) is 341-0638-03-A0. No GPU installed. We use 230V only. I have already opened the TAC case and it looks like it will be escalated. I just want to know if anybody else has the same issue like we have here. 

Version: 4.0(2c)

[Kirk J]

Any chance you could borrow the dual 1600 watt PSUs from the other unit temporarily to test?

Testing the 'problem' unit with different model PSUs, might help rule in/out if the unit has a problem power backplane, vs some odd firmware issue.  If you have the same issue with the 1600watt PSUs, then I would tend to suspect power backplane.

I checked our ticket system, and don't see other cases besides yours with this problem, so it may just be one off hardware issue.

Kirk...

[peter.matuska1]

hi, sorry for the late response. Tested with the 1600W and the issue is the same. We replaced the server via RMA and the new server is fine. No issue there.

[Kirk J]

Glad to hear it is resolved.

Sounds like something with the power backplane then, which would have been replaced when the overall server chassis was replaced.

Kirk..