DL360 G3 - amber interlock light / bad internal health -- help

[lyle scott]

I have tried everything posted on these forums and anything I can find on google to set my server straight.

Basically, the interlock light on the mobo is amber and the bad internal health light it red on the front.

I am using hotpluggable power supplies, so I put flipped switch 4 on SW1.

I have reseated everything on the board.. and taken out anything I can get to at one point or another.

The processor does not get hot and i dont think the harddrives are spinning, either. The power supply makes the fans whirl.. that is about it.

The main reason for posting: how can I track down what component it is? multimeter? I can't get docs on the debug lights that the manual mentions.

I would like to replace the part asap, i just dont know what to do! :) help!

[Matti_Kurkela]

"Interlock" usually means a system that stops the server from starting unless all the critical components are in place.

In a 1U slim-line server, the correct flow of cooling air is very critical. Are all the fan modules in place?

Even removing the top panel of the server might be enough to disrupt the internal airflows so that the processor won't get properly cooled. In such a case, the server designers usually equip the chassis with an interlock switch that prevents the server from starting if the chassis is open.

Also make sure that all the connectors between the circuit boards of the server are *fully* mated.

MK

[lyle scott]

Should have mention I had tried that, too. I have another server just like this, and it works with the case open.

All fan modules are in place.

I just basically want to know how to troubleshoot for a bad component. How can i debug to find the bad component?

[lyle scott]

update:

actually I just noticed my CPU fans are not coming on. Is this because of another failure, no reason for them to come on.

Or could this actually be the problem?

These are the fans that are behind the power buttons on the front of the case.

[Rob Buxton]

Not quite sure from the post - do you have a second power supply in the server/
These can support two x power supplies.

If you have another DL360 G3 You could try swapping power supplies. Sounds like that might be dead.

I've had issues with power supplies on DL360s before.

[lyle scott]

I actually have 2 power supplies. By experimenting, they both work (I assume?). They each power up the motherboard (no post, anything on monitor). I can keep both plugged in and hotswap them/unplug them and everything is fine.

The green lights don't come on in the back.. but the power is definitely getting somewhere. Everything seems to be "on" except the CPU fans and I don't think the hdd's are spinning. Also, I can not turn the server on.. as soon as i plug it in, the thing comes on.
javascript:postAnswerSubmit('submit');
I don't have access to the other one :-/

You'd think they would have some indication on what component is failing... or atleast documentation on the debug lights they flaunt in the manual. Never again...

[cnb]

Debug switches and code info on Page 152:

http://bizsupport1.austin.hp.com/bc/docs/support/SupportManual/c00300504/c00300504.pdf

HTH,

[lyle scott]

awesome..i like the guide. I don't get where the port 85 display codes are..or how to make sure that it is set.

I see 8 tiny LEDs when I throw sw3 in debug mode. They all come on.

ALso, the green power light does not come on in the back of my powersupplies (both!). Is this bad?

[cnb]

Use the maintenance and service guide to set-up the switches.

http://bizsupport2.austin.hp.com/bc/docs/support/SupportManual/c01125247/c01125247.pdf

There's also an LED table to help isolate.

HTH,

[lyle scott]

ah.. i swore i tried this last night, but the lights are actually working blinking. I get some action on the top 2 iLO diagnostic LEDs... (farthest from power supply). I think they are labeled cr25 and cr26 ... they are both blinking on and off at steady (right under a sec) intervals (not in sync though)

[lyle scott]

or PAL LED 6 and 7 .. i think that is what they are called.

[cnb]

Lyle,

What is the server part number (xxxxxx-xxx) and was this ever working? What did the POST 85 codes indicate?

[lyle scott]

cnb: hm.. do you mean mother board part number? 305439-001 is the mobo...

It was working before i had it shipped to my new address. i had just booted it up for the first time in a few months when i originally posted.


http://forums.itrc.hp.com/service/forums/questionanswer.do?threadId=1300392

^^ is a {desperate} update. I'm not sure what 85 codes you are referring to... my board doesnt POST.. the harddrives, cpu fans, optical drives, proc dont turn on when i hit the power button..

[cnb · Accepted solution]

Lyle,

Not exactly. There is a part number also for the server (should be on the bezel). There are several configurations for this server. The manuals I referred to contain a section for setting the switches when you can't POST. They check the basic funtionality to isolate 'pre-POST' flow failures.

Use the maintenance and service guide to set-up the switches.

http://bizsupport2.austin.hp.com/bc/docs/support/SupportManual/c01125247/c01125247.pdf

There's also an LED table to help isolate.


Debug switches and code info on Page 152:

http://bizsupport1.austin.hp.com/bc/docs/support/SupportManual/c00300504/c00300504.pdf

If you can't get the 84/85 procedure to work, you may have to just replace the m/b at this point. Interlock failures are 'usually' easy to resolve (PCI riser, cover switches, etc...). However if you've eliminateded all possible interlock ares then maybe it's an internal m/b failure.

There are other threads on this forum with this part number and failing to POST. Their only option was to replace the M/B.

http://forums13.itrc.hp.com/service/forums/questionanswer.do?admit=109447627+1230842555069+28353475&threadId=1144758

ls> awesome..i like the guide.

Also see this thread for those who may also have simliar issues to aid them in problem resolutiom:

Please read the following about assigning points:
http://forums.itrc.hp.com/service/forums/helptips.do?#33

HTH,

[cnb]

Clarification:

That s/b PORT 84/85 codes. Sorry for the confusion.

[lyle scott]

ah..i think this is the serial: M00TLGP11R (that is on the case, under the cover, right above bezel)

As far as the port 84 codes... I had tried setting 1,2 to on,on and 1,2 to off,off which is what the manual says to do for the code. I am not getting any LED lights that way... isn't this the way to read the 84/85 code? (I had read interpret the blinks at bits to convert to hex)

the top 2 LEDs blink when i put the switch to iLo debug.. but when i go to the ILo web control panel, it doesnt tell me anything.

..am I wrong about how to read the 84/85 code?

[cnb]

See this thread:

http://forums13.itrc.hp.com/service/forums/questionanswer.do?admit=109447627+1230938860292+28353475&threadId=1262637

HTH,

[cnb]

PORT 84 errors *should* be logged in iLO:

System Diagnostics - Integrated Lights-Out may be used to diagnose systems. The Remote Console may be used to monitor the system for POST error messages, the Integrated Management Log and iLO Event Log record events useful for diagnostics, Virtual Floppy (if licensed) may be used to boot and run System Diagnostics. A new feature with Integrated Lights-Out is the ability to record server â Port 84 POST Codes�? as the system boots. These codes document the progress of the server through the booting process.

http://h20000.www2.hp.com/bizsupport/TechSupport/Document.jsp?lang=en&cc=us&taskId=120&prodSeriesId=332279&prodTypeId=18964&objectID=c00333965

HTH,

[cnb]

Lyle,

Unplug AC

Remove system access top cover.

Check that PPM modules are installed for each CPU.

Only 1 power suppy installed (right one as viewed from the rear)

Set your system board switches to:

SW1 2ON, all others off
SW2 3ON, all others off
SW3 ALL OFF

Connect VGA Monitor, KB and Mouse directly to the system.

Connect a known good network cable to iLO port.

Plug AC cord in. At this point 3 things will happen:

Power switch will glow orange.

DC fan assembly (x3 fans at power supply interconnect) will turn on and spin.

LED CR2 on System board (near CD/FDD connector) will light, indicating System Board (SB) power.

Note: the Green Power Supply LED in rear on the Primary P/S will NOT be on at this point.

All other LEDS on System Board (SB) will be off.

If INTERLOCK is on then one of the PCI riser assemblies is either faulty, not seated properly or there is dust/debris in the PCI Riser connector causing a poor connection. You must have BOTH PCI Riser assemblies installed and functioning (even if you're not using any PCI adapters).

If the ILO port IP address and subnet are on your local network and you know the IP address, then you should be able to browse to it using a web browser and login. Don't use telnet.

Please report your findings.

HTH,

[lyle scott]

Thank you for your great reply.

Sill nothing in iLo (everything reports as "empty" and iLo debug is all at "PASS".

I don't have anything in my PCI riser cards. They are both blink. (The two white slots parallel to power supply).

My symptoms are as you describe though... to the T.

[lyle scott]

and.. by calculating the PAL leds to binary:
1100000 == 0xC0 which.. not sure how to relate that to page 35 in troubleshooting guide (table of codes).

[lyle scott]

with the lights like you have them, all the lights turn solid green. Aren't these the SmartArray 5i diagnostic LEDs?

[cnb]

Hmmm...there should not be any decode/port LEDS at this point since we haven't energized the DC power switch. So I'm confused on why/how you are getting the decode LEDS, since a) no DC power has been applied to the system to initiate POST and b) the switches should not be in decode mode.

Did you push (energize) the POWER switch after plugging in the AC?

Do you still have the INTERLOCK LED with system in the powered off state with AC plugged in?

[lyle scott]

The power button does nothing in any case... it's always just amber.

The interlock light is always on in any case as well. It never changes... you can pull all hardware you want, it is still on. No other lights have ever come on on the board... except the one by the hinge that the cover connects to (CR2).

Other than pulling critical power points, I can remove components (CPU, PPM, RAM, HDs, etc) and there is not a noticeable difference in symptoms (even debug lights stay the same).

FYI:
(1) 1 processor... and PPM is in correct slot.
(2) still nothing in iLo.. it just says the system is off. No post, etc in virtual terminal
(3) Debug 0000: no lights
Debug 1000: top 2 blink (6-7 on 0-7 scale)
Debug 0100: all lights solid

hm.. i'm doubting this mobo :-/

[cnb]

Lyle,

I need you clarify something in your earlier post, you indicate that both System Board PCI riser assembly slots are empty. Is that true? If it is, then this system won't ever pass POST. The system requires the risers whether or not you populate them.

If you do have the riser cards then:

Remove the two PCI riser assemblies, take a gold wipe or alcopad (90%) and clean the connector edges on each, let DRY completely. Use a flashlight and inspect inside the PCI riser slots on the SB. Make sure that there's no debris or dust by using a can of compressed air and blow out the slot sockets.

You won't get anywhere unless INTERLOCK is off. Then reinstall and/or swap positions with each other. The INTERLOCK indicates faulty riser connection issues.

HTH,