Supermicro X8DTH-iF and 6F, no Post, no Beep, working IPMI

[jakobgeb]

Hey Guys,

i got the "same" problem with two different Supermicro boards and i am out of ideas whats the problem. So i got the X8DTH-6F, two Xeon X5650 and 6x 8GB ECC 1333Mhz Samsung memory. Assemlby, i got a post right away without VGA connected. Putting it in the the chassis, post and no problems, but no video signal with VGA connected. Then powering it off, disconnecting power to switch some cabels and suddenly no post and no beeps. taking out and of everything but CPU and RAM, nothing but a working BMC, so the IPMI is working. then some different combinations of one CPU on RAM stick, nothing. also CMOS reset per jumper and with taking out the battery, only the BMC LED is indicating a working status.

So i assumed, that i somehow fried the board...

Second try with the X8DTH-6F, to ensure a working system assembly again on a antistatic base with working post and VGA signal (same CPU, RAM, PSU, display as before). At this point i can be sure that CPU and RAM is fine, can i? Then putting it in the chassis, booting into bios, configuring IMPI (unfortunatly i forgot to change the pass) and then booting into Ubuntu, all fine. Powering down, disconnecting power to boot it up into bios again to alter the IPMI pass and DAFUG i doesn't post anymore.

I am really out of ideas what causes the problem. I kinda think that this is some kind of bios issue, because the hardware should be fine after around 5 boot ups. Also why should two board fail in a similar scenario? can a working cpu short out a board? i rule out the case for a shorting problem, because i was able to boot up the system.

I would really appreciate some thoughts to this...

With kindly Regards
Jakob

 

[Fritz]

PSU ?

Rogue MB standoff in chassis?

 

[jakobgeb]

Sorry, PowerSupplyUnit...
No already checkt the standoffs :-/

 

[Fritz]

Could be a bad MB. I suspect I may have a X8DTE-F that has screwed the pooch. Gonna pull it out next week and test it.

Good luck hope it turns out to be something simple.

 

[Terry Kennedy]

I am really out of ideas what causes the problem. I kinda think that this is some kind of bios issue, because the hardware should be fine after around 5 boot ups. Also why should two board fail in a similar scenario? can a working cpu short out a board? i rule out the case for a shorting problem, because i was able to boot up the system.

I would really appreciate some thoughts to this...

Those are some pretty solid motherboards (I'm running a few dozen of them).

The first thing to do is to get back to a minimum configuration and see what happens. Make sure you have both of the EPS12V connectors attached, even if running only one CPU. For testing with a single CPU and a single stick of memory, a "cheater cord" should work if you don't have a power supply with a pair of EPS12V connections.

Have you tried a different power supply? I find it handy to have a known-good bench supply (I use a pair of older, non-SQ Supermicro 920W units and a spare power supply backplate/cables).

Put a processor in the CPU 1 socket and a stick of compatible memory in the P1DIMM1A socket. At least with registered memory, the board is pretty flexible as to what socket you actually put the memory in.

To rule out the onboard VGA, if you have an add-in PCIe video card you could try using that. The built-in IPMI remote console won't work with an add-in card, but you should at least be able to see if the system POSTS using the add-in card. If you have a monitor where you can disable power-save, that would also be helpful - messages that flash on the screen briefly may be missed if the monitor is in power save at the time.

I've never run into a CPU this board doesn't support with even an old BIOS, but an E5520 is a convenient chip to keep around for testing purposes.

If all else fails, you could try a PCI Express "port 80" diagnostic card.

If you're in the US, I'd be glad to look at one / both boards for you and see if I can get them going, if you get completely stuck. It isn't practical to do this if you're outside the US due to shipping costs, customs forms, etc. but perhaps in that case there's someone near you who could help.

 

[jakobgeb]

@Fritz: yea could be, but two models in a row with the same problem... i hope so to, thx

@Terry Kennedy: i already tried one cpu in both sockets with one dimm. also got to two eps12v. and also tried another power supply, but also not a "server" power supply. i will try a supermicro power supply, known to work with this model.

the add-in VGA was no help either, because the board doesn't even power on... the onboard VGA from the second board was already working until the board stopped powering up.

good idea with the diagnostic card, will try this tomorrow. but would i get some codes from it if the board does not power up ??

really thx for the offer, but germany is a little bit to far away ^^ really appreciate it...

i read some threads about supermicro boards beeing picky on ther choice of power supply, so i hope the supermicro power supply does the trick...

 

[JoeyLovesTrains]

@Terry Kennedy I have a similar Problem with my X8DTH-6F, It will power on and all is good both ethernet jacks are indicating activity, but the USB ports and VGA aren't doing anything. The BMC is blinking and there aren't any warning beeps. I've also tried clearing CMOS and removing the battery, but nothing's happening. It'll start up and the reset/power buttons are responsive and everything seems to be working except most of the rear io.

 

[sfbayzfs]

The north bridges on the x8 dual Xeon boards tend to run hot, and the models with an extra chip for extra PCIe lanes like the X8DTH-6F seems to get permanent heat damage if you don't cool them enough - I know myself and at least two other forum members who haven't posted in this thread yet have had that particular board die. The only issue I ever has with a single Northbridge heatsink X8 board was n X8DTE-F which was my own fault (sad story - I was checking and flashing the BIOS on a recently acquired board and didn't notice that one corner of the heatsink wasn't screwed down, so the CPU overheated and malfunctioned while flashing the BIOS, I foolishly retried and it failed worse, and then it never booted again.)