Over Heating

[mdasher]

According to the management console, every night, at around 10:00pm, the server cpu reaches 81 degrees Celcius. A few warnings are sent and then after 10 minutes the server shuts down. The over heating warnings also occur during POST, even if the server has been off for 10 minutes +. This has been an ongoing problem over the past few months. If the warnings were true the CPU would be fried by now. The obvious answer is that a sensor is bad. Where, what, and how do I remove/replace the sensor?

[Matti_Kurkela]

In many recent CPUs, the temperature sensor is located inside the actual CPU, so replacing it would mean replacing the entire CPU.

Even if the sensor is a separate component, it is unlikely to be field-replaceable. If you don't already have experience in hand-soldering surface-mount components with tiny connections, forget it.

There is a function called "thermal throttling" (at least) in all current Intel CPUs: if the CPU cooling has failed, the CPU will severely limit its speed to prevent damage. It will still get very hot, but not hot enough to cause immediate destruction.

What's the model of the server?

[Bruno Mesquita_1]

If the Server is a Proliant ML350 G4 3.02Ghz, Xeon, is probably the Heatsink that should be changed. Spare Part 366866-001.

After replace if the problem remains you should consider move the server to a cooler location.

Regards

[BPatrick]

Hi,
I had the same problem with one of our DL380 servers. HP support has asked me to reseat the CPU and after that I never got this problem. So you could try that.
Also you could check all the fans are working and also the heat sink is properly fitted.

Regards
Patrick

[Proliant BL,DL ML]

Hi Mdasher
Have had this issue earlier and found that the processor heatsink with fan has been installed backwards.
As fix you can try to remove and reinstall the processor heatsink assembly so that the fan is on the Media bay side of the processor (180 degree turn).
If this doesnt hold any good or doesnt help.

The thermal sensor is on the system board and would request you to log a call with HP tech.support since if the sensor is causing issues the system brd needs replacement.

Regards
Riju

[Liquid_Void]

Hi mdasher, If you can specify which server this is, i think i have a solution for you.

Kind regards

Riaan

[mdasher]

Yesterday I removed the CPU and heatsink. I think the problem was actually the amount of Thermal compound or the lack there of. Also, the thermal compound that was on there was dry and flakie. I added a ton of thermal compound, reseated the heatsink and the problem seems to be fixed. Thank you all for the great responces.