Unable to power on server via ILO or physically press the ON button

[Geraldo45]

Hi All

I have a srange situation where a remote server crashes unexpectedly and I am unable to power on the server using ILO or get the local support staff to power it on by pressing the power button.  The only way I can power it on, is to get them to disconnect the power cables and reconnect them which turns on the server.

However if I cleanly shutdown the server, I can power it on by ILO.  So it must be something to do with the way the server crashes which causes the problem of not being able to power on. 

The crash seems to be random, it is not a BSOD and there are no ASR events in the IML, the server has not been changed in anyway and thre are no errors in HP SMH or Online Diagnostics.

I am thinking of replacing the system board but was wondering if anyone had any advice that may help me?

Thanks

[Suman_1978]

Hi,

You forgot to mention the server model and operating system version.

Before replacing the system board, do check the following:

See if any BIOS update fixes this issue.

Swap with another power supply and check if it crashes, this is because you are removing the power to the server completely to get it back online.

Thank You!
I am a HP employee.

Useful Links for ProLiant Servers

[Moderator edit: Removed the broken link. Please refer to https://support.hpe.com/]

[Geraldo45]

Forgot to add the crucial detail its  a DL 360 G7 with W2K8 R2 installed.  I've noticed in the ILO logs a pattern of events each time this happens, in this order

Server Reset

Server power removed

The above events are logged immediately, I can see Power-on signal events sent by whoever tried to power on the server remotely and then when the on site person powers the server on by reconnecting the power lead and then pressing the power button there is a Power Restored to ILO event.

We will try and replace the PSU(s) first

[Rob Twiggs]

I've also got a DL360 G7 that was shutting down every 20 minutes or so with the same iLO event log entries on each unexpected shutdown: "Server power removed". No other diagnostic errors were ever logged. System Health LED would also flash red on each unexpected shutdown.

Fix, in my case, was to remove the two hot-plug power supplies and rub the hot-plug contacts on each power supply with an eraser to clean them.

Eraser can most probably also be used with some success on memory stick contacts, PCIE connectors, SATA connectors, etc when troubleshooting Proliant Servers.

[sirkro]

The most likely reason why you cannot bring up the server again via the power on button or ILO is that it went down due to a Voltage rail failure. This is why you are able to power it up normally after a normal shutdown.

SO some component causes a Voltage rail to go down, replacing systemboard might be a good start but unless any diagnostic , LED status or HW log file (i.e. IML) indicates a voltage regulator fault it will be trial and error unfortunately.

HTH

Kris