DL380-G6 overtemperature led

[caesar808]

Hy everyone. I have a problem and i hope that some of you can help me.

I have one Proliant DL380 g6. One morning the server has crashed and when i try to switch on, nothing happened.

The server is try to start, but after a moment (less than a second) the server is stop and amber led OverTemp is on.

The machine doing this passage as in a loop an try to swtich on avery 30 second approsimatively.

I try to remove fan, and move position, try to change the PSU (i have 2 PSU) that is working, but the problem persist, i try to unplug all (SCSI bay, CDROM, leave only 1 ram etcc) but i can't solve the problem.

Any ideas?

Thanks everyone. And i'm apologize now for my errors.

[Suman_1978]

Hi,

Have you tried to clear NVRAM from the System Maintenance Switch?

Do check "Fan Failure, Fan Solution Not Sufficient, Temperature Violation Detected, 1610, 1611 Errors at POST".

Do check "Server May on Rare Occasions Stop Responding During Power-On Self-Test (POST)"

Thank You!
I am a HP employee.

[caesar808]

Hi,

first of all thanks for your answer and very fast help.

About the first solutions, i already tried to clean the fans and change the fans position. I don't try the firmware upgrade yet, i upgraded only the iLo2 firmware (v2.23 actually).

For more details, i checked the iLo2 log, and message "System switch off due to Overtemperature" is repeated.

Sometimes (every two message of temperature) i see "Brown-out recovery" but i don't know what does it mean.

Do you have any information about upgrade the firmware (maybe the procedure and the link to download the updated version)?

During this evening i will try to clear CMOS (i don't remember if i already tried) and i let you know.

Thanks Again!

[Suman_1978]

Hi,

Brown-Out recovery
A brown-out condition occurs when power momentarily is lost to the server. A brown-out interrupts the operating system, but does not interrupt the iLO firmware. Under brown-out conditions the iLO service remains uninterrupted for about 4 seconds (longer power interruptions result in black-outs).
Support has been added to iLO to detect and recover from power brown-outs. If iLO detects that a brown-out has occurred, server power is restored after the power-on delay. After the brown-out recovery, iLO firmware records a Brown-out recovery event in the event log.

Here is the BIOS update using Windows OS.

Here is the BIOS update using USB Key in case server is unable to boot into OS.

Thank You!
I am a HP employee.

[TVVJ]

Check if the power source does not have loose connections as brown-outs suggests interruptions for a very  brief intervals.

Also check if the processor heatsink is installed correctly. Without removing the processor, re-apply the thermal past.

Regards,

Vijayasarathy 

[caesar808]

Hi,

i tried all the solutions that you have suggested, i clear the CMOS using maintenance switch (6) and nothing change.

I tried to remove the heatsink of CPU but the message is the same: "Power loss due to overheating. Attempting to restore power"

How can i upgrade the firmware without any input and without boot?

Thanks again.

[Suman_1978]

Hi,

If this is the message "Power loss due to CPU overheating.  Attempting to restore power" then the issue could be with Processor itself of system board.

Try to swap with another processor if you have.

Thank You!
I am a HP employee.

[Oscar A. Perez]

When a temperature sensor on the system board raises the Temp-Deadly signal to the system PAL (CPLD),  the PAL forces iLO to shutdown the server immediately. No questions asked.

These are the 3 possible messages logged by iLO2 during these temp-deadly events:

"Power loss due to CPU overheating. Attempting to restore power"
"Power loss due to chipset overheating. Attempting to restore power"

"Power loss due to overheating. Attempting to restore power"

The first two messages are self-explanatory. The third one is tricky to root cause since there are several temp sensors on the system board. Try using the iLO2 webIU to pin point the zone that is overheating. Look under "System Information->Temperatures" while the server is powered On.  

If this is happening right after the server is powered on, it is most likely caused by a faulty sensor or some other hardware problem that could cause the temp-deadly signal to get asserted or stuck.

[caesar808]

First of all, thanks again.

Unfortunately clearing cmos made the iLo log clear, and on the system temperature i only see the message "server is switched off".

But i can check every temperature sensor with a tester, and search where is the faulty sensor.
Anyone have a data-sheet o schemas about positions of all the motherboard temperature sensor?

And when i identified the sensor that cause the signal, how can i remove, repair or bypass?

[Oscar A. Perez]

Bypassing a sensor would be a bad idea.  

If you cannot just replace the system board and really want to repair it, look for the SMSC EMC14xx chips on the board and check each of the remote diode sensors attached to them in order to figure out which one is causing the themal shutdown.