DL980 G7 E7 not starting

[Jelle1]

Hello,

I have an HP DL980 G7 E7 which is not starting.

It starts 4 times for about 2 seconds, then flashes red health on the SID.

It is running with 8 E7 4870 CPUs, all of them verified to work in another server.

It has 64x 16GB RAM, all HP branded and verified to be working in another server.

It also has amber LED on XNC. The XNC has been replaced with a working one but still I get the error.

The XNC cables have been verified to be correct 3 times.

I don't know iLO IP of the server to get details.

I tried removing the battery from the SPI board for 10 minutes and booting up again

I tried removing the XNC board and lower CPU tray (Only running 4 CPUs) I still get XNC error

I tried pulling out all expansion cards (Controller, SSDs, HBA cards)

The LED on the systemboard is writing 80 and 00, changing each second

What else can I try?

The attached picture shows the systemboard. The 3 amber LEDS that's on only turns on right after the server failed to start. When the server tries to start these amber are not lit.

Thanks for your help!

[FabioC1]

hI  Jelle1

Do you tried minimal confuguration, 2 memories connected per  processor, only ?

https://support.hpe.com/hpsc/doc/public/display?sp4ts.oid=4231377&docLocale=en_US&docId=emr_na-c02510072

Aut of equipment have some led blink ?

https://support.hpe.com/hpsc/doc/public/display?docId=emr_na-c02524679

[Jelle1]

Hello FabioC1

I just tried minimal configuration, exactly same result.

For the fun of it I tried no memory cartriges at all, again same result.

If I start it without upper or lower shelf the power button is blinking green.

Health LED is blinking red after 4 failed starting attempts

Power LED is amber

XNC LED is amber

Thank you for your reply and help thus far.

I do not have a replacement systemboard I can try with, My 4 other DL980 G7 are not E7 compatible.

Since it does this even without RAM installed at all, it does point towards the system board for me at least.

[FabioC1]

 Jelle1 I think that SPI Board is fault.

Red (flashing)

Internal system health critical

https://support.hpe.com/hpsc/doc/public/display?docId=emr_na-c02498130

System Peripheral Interface (SPI) board - Includes the rear panel NC375i NIC (4x), iLO3, USB, serial port, VGA, keyboard and mouse connectors, plus internal mini-SAS connectors, SD card slot, and the system RTC battery

AM426-69017

AM451-69002

Man

If equipment has warranty will not necessary pay for replacement.

Can you check warranty in this site. - https://support.hpe.com/hpsc/wc/public/home

If the equipment is out of warranty you need to pay .

Can I call to HPE Support and ask for replacement this part

https://pg-receiver-pro.glb.itcs.hpe.com/WclWeb/WclEntry.aspx?origin=acc&cc=US&lang=en&co=hpe

[Jelle1]

Hello again FabioC1,

I replaced the SPI board with one from a working server (Same sparepart) but the server still just tries to boot 4 times before giving up and flashing red health LED and amber XNC (Even with no XNC installed and no lower drawer installed)

There is no longer warranty on the server, and I do not have the money to pay HP to look at it. I think better is to get a full replacement server somewhere, although that is expansive too.

Sadly I do not have that systemboard in another server to try with, the other 4 servers I have are with X7560 CPUs and old sparepart systemboards.

I just tried to throw in the XNC and lower drawer, and with the other SPI board. Still same issue. The SID has amber LED in processor 1, 2, 3 and 4 in lower drawer along with XNC and cabling. These LEDs are lit even before trying to turn on the system for the first time after restoring power to it.

[toothpick]

I am having the EXACT same problem with my DL580 G7. No video output. Same lights on system board, Soon as power is connected, the Port 85 display flashes between 80 and 00 for about a minute (i assume until the SPI has booted) then it switches to a perminant 00.  The SID on the front has solid amber lights on CPU 1,2,3,4 (all of them) the health light is flashing red. I have tried all steps mentioned so far in this thread with no results. The CPU's never seem to get warm to the touch yet the chipset chips do get warm quickly.

I decided to leave it running like that with fans at full speed for about an hour. To my suprise, after 1+ hours, the fans slowed to a normal sounding speed, yet no other difference... I thought it had just timed out and shut off so I walked away. 15 minutes later I noticed a electric sounding click every minute coming from the server. I investigated and instantly burned my hand on the CPU cover! All four were at least 100c tempature yet the fans were completely OFF! I yanked the power cord and reattached to get them spinning again, but as soon as they spun up, they spun down and completely stopped again... The fans 12v power is attached directly to the Power board while the PWM speed control signals attached to the SPI board. So I yanked just the PWM cable off the SPI and the fans instantly kicked into full speed.

So now, the fans will not operate while attached to SPI.... Clearly the SPI is faulty somehow right? So I tried connection to ILO, which seems to work totally fine and shows no IML or any error information. It shows the CPUs, memory and fans in real time. Power on/off controls work as well as the power button. The only thing odd I see is most the actual stats (fan rpm, temps, power data...) just say "unknown" or "unavailable"

So that blows away every idea I could come up with! I have no idea what to even start replacing here but I have to get it working somehow... Im really hoping someone here has figured this out! After finally finding this thread, I know im not the only one, so that gives me some hope.   TIA!!!

[raellear]

Do you have any news on this issue? I have now two machines with the same problem lying around here, see the following thread: https://community.hpe.com/t5/ProLiant-Servers-ML-DL-SL/DL580-G7-not-posting/m-p/7015912

I could ask the shop that sould me these to replace them... If it really is just a motherboard issue, I am not sure! If we only knew the cause...

[raellear]

Any update on this? I experience the same problem with two dl580 G7. it realle seems to be a curse. Look at my thread: https://community.hpe.com/t5/ProLiant-Servers-ML-DL-SL/DL580-G7-not-posting/m-p/7015912

[Giuseppe1]

I have a DL 980 G7 that does not start (no display no POST). At power up the first 3 LEDs on the left side of the System Insight Display are green (power, boot and heartbeat; boot is flashing green). After 2 min the led boot flashing amber, the display remains black. The system board has 1 green LED and 2 amber LEDs on the right edge (next to Slot 7) , the SPI board has 2 green LEDs on, the SPI board's ethernet sockets, if connected, flashes green. Can someone help me?

[bennyb]

DL980 G7 Systems Fail the Drawers, SPI and XNC module most frequently.  The main issue in replacing them is to make sure they all have the same firmware bundle on them. The firmware complex for these requires that the levels are all part of the same bundle on the 3 main components, the  SPI, XNC and CPU drawers although there is some microcode on the PCI Board as well.

We stock our tested spare parts based on the firmware bundle from the system they were tested in so we know that the XNC for example, has the correct version of firmware on it so that it will work in a system with a specific bundle.  If there is any mismatch you will get these errors repetitively.  We have also noticed that many self-maintainers or engineers with little experience with these systems tend to bend CPU socket pins which makes it look like all sorts of funny things are happening.  If someone uncertified has ever removed CPU's please check the pins and save yourself some time.  Also recommend to check the winchester connectors on any component that has been or will be replaced as they or the pins they connect to can be bent easily and cause these issues that seem to have no logical sense behind them.