BL685c G6 power issue. Need advice.

[chuckk281]

Pierre had a customer issue with powering up a chassis:

***************************

Hi

This situation is with a new blades enclosure installation. I have a power issue with BL685 G6, when I do power up the server the server will start to boot, but I get the following error message at the console. “Enclosure Power Event Detected”, “System Halted until Power condition is corrected”. (See attachment power.gif) In the OA log (see oa.log) there is no error message. The c-7000 display panel is all green, no error. If I take the blade out of the bay, reseat the blade and power the blade it boot ok. I saw this error on the first day of installation, did not pay that much attention because the blade was ok after reseating the blade. We did flash the enclosure and blades with Firmware CD 9.1c.

On the second day when trying to use RDP to push a OS, we did power the blade again, the issue came back. If I reseat the blade, it will boot ok. I do not know if this will reoccur again, I need to test.

The configuration of the enclosure is a c7000, with 8 BL685 G6, two Flex-10 modules and two OA. The enclosure have 6 power supply but only one side (3 power supply) do get power, the customer do not have extra power in PDU to connect the other 3 power supply. I do not the information on how the OA is configure for power redundancy. Does someone have seen this kind of error? Having 3 power supplies should be ok to power all the 8 blades but will not provide any redundancy. This error is present event when only one server is power up.

Any inputs on what to look for to resolve the issue?

********************************

Hoa joined the discussion:

****************************************

I don’t think you have enough power to run the enclosure with three power supplies in place without errors.      Below links may help you to understand more on the issue.     Try to enable Dynamic Power, enable AC Redundancy via front panel or OA User Guide instruction.

• Blades System Power Sizer: http://h71019.www7.hp.com/ActiveAnswers/cache/347628-0-0-0-121.html

• Power Advisory: http://h18000.www1.hp.com/products/solutions/power/advisor-online/HPPowerAdvisor.html

• OA User Guide http://h20000.www2.hp.com/bc/docs/support/SupportManual/c00705292/c00705292.pdf

****************************

What do you think? Any other thoughts on this?

[keithnj]

We have the same intermittent problem.  Here's our configuration and don't think we are oversubscribed to power.

Bay	Device Model	Firmware Component	Current Version
1	ProLiant BL460c G6	System ROM	I24 03/30/2010
		iLO2	1.81 Jan 15 2010
		Power Management Controller	3.4
2	ProLiant BL460c G6	System ROM	I24 03/30/2010
		iLO2	1.81 Jan 15 2010
		Power Management Controller	3.4
3	ProLiant BL460c G6	System ROM	I24 03/30/2010
		iLO2	iLO2  1.81 Jan 15 2010
		Power Management Controller	3.4
4	ProLiant BL460c G6	System ROM	I24 03/30/2010
		iLO2	1.81 Jan 15 2010
		Power Management Controller	3.4

Bay	Device Model	Firmware Version
1	HP 1/10Gb VC-Enet Module	2.32

Interconnect Firmware Information

Device Name	BladeSystem c7000 Onboard Administrator
Manufacturer	HP
Firmware Version	3.11 Aug 19 2010
Hardware Version	B1
Part Number	412142-B21
Spare Part Number	414055-001
Serial Number	…...

Enclosure Ambient Temperature	21°C / 69.8°F
Thermal Subsystem Status	OK
Power Subsystem Status	OK
Power Mode	Not Redundant
Present Power	1096 Watts AC
Power Limit	10944 Watts AC

There was a “degraded” notice on the OA screen for the chassis regarding the power supplies in bays 4 and 6 but this notice has been there all along without issue.  Power bays 4 and 6 are empty.

I just checked the blades in NJ and they all rebooted and now they are stuck at the HP screen reading “Enclosure Power Event Detected. System Halted until Power Condition is corrected”.

Server OS config:  Vmware ESXi 4.0.0 Realeasebuild-261974

Server Hardware config:  ProLiant BL460c G6, 2x X5650 @ 2.67GHz, 96 GB

4 blades installed.

6 power supplies installed.

[Mav83]

The servers need to be power cycled to get rid of this error. Either do it from the ilo config page or you could take he servers out of the bay and plug them back in to power cycle them. This should resolve the error.