X10SL7-F failure?

[sremme]

I apologize because I now this is slightly off-topic but this forum seems to be the most knowledgeable given the hardware I have.

I have had a server running 3+ yrs with a X10SL7-F and G3258. Yesterday I rebooted the server and it won't power back on. When I look at the motherboard and lan ports .... there are no leds on at all. I tested the power supply and it was good, but I replaced it anyway and still no leds light up and the system won't power on. I replaced the cmos battery (just because) and removed all the power cables except the 24-pin and 8-pin cpu. Still no leds and won't power on.

So a few questions:
Is there anything else I could try, for example resetting the cmos, to see if I can get the leds on and system to power up?

Is there any way to check/test the motherboard before assuming it has just died?

Could the processor cause this? Gone bad or needs to be re-seated?

What is the easiest, and/or recommended, path to recover my server. Should I replace with the same motherboard ($250) to reuse my cpu and memory? Or would it be better, and maybe cheaper, to go with one of the other recommended builds on here? Which one(s)?


I don't have any current plans to expand beyond the 8x3TB drives I have now. But if the cost is reasonable future upgradability might be nice.


Thank you for any advice, opinions, and recommendations.

 

[l@e]

X10 is not so old to die of age. I have pleanty of them even x9 and some x8 that are 24/7 from years.
The only thing that comes up my mind that a mobo will not start up if it was in operation, and you did a restart manualy (not by some power failure or storm in the area) is a capacitor gone bad.
If you dont see anything suspicious on the board as a broken ic or blown capacitor you can try smth to force start the psu.
to not cause more problems it is better to disconect hdd and other peripheral power or unplug pci if any.
Then shorten the gren wire that is on the ATX connector with one of the black next to it. This way the psu will go live (you should hear the fan) and check the rest of the board for lights. If it boots normaly as long you keep the shorted pins together, there is strong possibility that is a capacitor gone bad. Anyway dont excluse the posibility of other components gone bad.
Any way if it will not boot this way at all then you know for sure is dead.

Also if it boots and the problem is one of the capacitors than you should worry also on what psu this board was working. Capacitors on server board have average life 5+ years and apart of very low chance of fabrication problem the psu if giving higher voltages /or high temp for long runs are the only real cause.

 

[Chris Moore]

I apologize because I now this is slightly off-topic but this forum seems to be the most knowledgeable given the hardware I have.

Is this a FreeNAS system?

Is there any way to check/test the motherboard before assuming it has just died?

Do you use the IPMI management interface? Is it doing anything?

Could the processor cause this? Gone bad or needs to be re-seated?

I have seen a couple instances where reseating the CPU made a system run again.

What is the easiest, and/or recommended, path to recover my server. Should I replace with the same motherboard ($250) to reuse my cpu and memory? Or would it be better, and maybe cheaper, to go with one of the other recommended builds on here? Which one(s)?

You might want to create your own thread instead of jumping into an existing one, but some additional testing is probably called for at this point.

 

[Ericloewe]

You might want to create your own thread instead of jumping into an existing one, but some additional testing is probably called for at this point.

I've fixed this.

@sremme, the first step here is to see what the IPMI log says.

 

[Stux]

Is there normally an IPMI power LED on the mobo on these systems? Is it pulsing?

Seems likely the mobo is just dead to me.

I think one of these boards should at least light up the IPMI, even if the ram sockets and cpu sockets are empty... Right?

 

[Ericloewe]

Is there normally an IPMI power LED on the mobo on these systems?

Almost certain there is a heartbeat LED. Both of my Supermicro boards have them.

 

[dvand]

Yes, I have this board and the ipmi led starts up when you have power.

 

[DrKK]

So, this particular board behaves in this fashion when the CPU is not recognized by the BIOS. Did you by any chance upgrade or downgrade the BIOS while the system was powered down (the G3258, for example, was not recognized by the original BIOSes that shipped with the X10)? I know it's a long shot. But this is precisely the behavior when the CPU is either not seated correctly, or the BIOS does not recognize the SKU.

Another possibility: I have encountered certain cases with the X10 series will play dead if it is not convinced there is an active CPU fan plugged in to either FAN1 or FANA. Could there be a disconnected fan.?

 

[southwow]

Test the wall outlet, sounds stupid but could you have tripped a breaker or GFCI?

Try an ATX power supply directly connected to the board to eliminate any redundant power supply board issues. If you don't have redundant (i.e. slide-out power supplies), please disregard this.

If you're receiving power, you should see the IPMI LED and the interface should be active. Even without CPU, you should be able to connect.

If IPMI is working, maybe the momentary power switch is bad? Try turning on the power in IPMI or shorting the power button pins or even using a switch from another spare case.

If none of this is applicable, you're probably looking at a new board. My X10 has been 24/7 for 3 years without as much as a hiccup, but failures do happen regardless of somebody else's luck with any given piece of hardware.

Make sure you let us know what your resolution is!

 

[Jailer]

Almost certain there is a heartbeat LED.

Two of them. One for IPMI and one for the on board SAS controller.

Sounds like a dead board to me.

 

[DrKK]

Didn't he say the IMPI LED lights up?

 

[Jailer]

Didn't he say the IMPI LED lights up?

Still no leds and won't power on.

He didn't specify.

 

[sremme]

Sorry.... I've been gone for a week. I will try and reply to everyone's questions...

 

[sremme]

>Is this a FreeNAS system?

No it's not FreeNAS. I asked the question here because the forums are much more active than the opensolaris forums I used to frequent.
I worked for Sun Microsystems and started with the original Opensolaris and migrated to OmniOS CE over the years.

>Do you use the IPMI management interface? Is it doing anything?

I do use it and it's completely dead. There are no leds on anywhere on the motherboard or connectors on the back.

>I have seen a couple instances where reseating the CPU made a system run again.

Maybe I'll try this as a last resort, but I'm thinking a dead motherboard is most likely the culprit.

>You might want to create your own thread instead of jumping into an existing one, but some additional testing is probably called for at this point.

 

[sremme]

>Is there normally an IPMI power LED on the mobo on these systems? Is it pulsing?
No, no lights at all.

>Seems likely the mobo is just dead to me.

>I think one of these boards should at least light up the IPMI, even if the ram sockets and cpu sockets are empty... Right?

 

[sremme]

>Test the wall outlet, sounds stupid but could you have tripped a breaker or GFCI?
Tried different outlet and cord... No difference.

>Try an ATX power supply directly connected to the board to eliminate any redundant power supply board issues. If you don't have redundant (i.e. >slide-out power supplies), please disregard this.
Tried brand new power supply.... No difference. (Also, old power supply tests good with tester.)

>If you're receiving power, you should see the IPMI LED and the interface should be active. Even without CPU, you should be able to connect.
No leds, can't connect.

>If IPMI is working, maybe the momentary power switch is bad? Try turning on the power in IPMI or shorting the power button pins or even using a >switch from another spare case.

>If none of this is applicable, you're probably looking at a new board. My X10 has been 24/7 for 3 years without as much as a hiccup, but failures do >happen regardless of somebody else's luck with any given piece of hardware.

>Make sure you let us know what your resolution is!

 

[sremme]

I did do one additional test....

I jumpered the pins on the ATX cable to force the power supply on. I hear a relay clicking on/off in the power supply, LE6 turns on red, then turns off, also LE3 turns on green then off. The leds are cycling with the relay in the power supply. When I switch the power supply off using the rocker switch on the power supply, the case and cpu fan spin about a half revolution.

So my next questions are.... (Assuming the motherboard is dead)

Can I send the motherboard off for repair, or is it cheaper to just replace?

What would people recommend if I need to replace the motherboard? The easiest path of recovery would seem to be to replace with the same motherboard. But are there better solutions/boards today. I have no current plans to upgrade my server (I'm using 6x3TB drives and have two connectors available), but maybe I should allow for ease of upgrades in the future? Recommendations appreciated. :)

 

[sremme]

>So, this particular board behaves in this fashion when the CPU is not recognized by the BIOS. Did you by any chance upgrade or downgrade the BIOS while the system was powered down (the G3258, for example, was not recognized by the original BIOSes that shipped with the X10)? I know it's a long shot. But this is precisely the behavior when the CPU is either not seated correctly, or the BIOS does not recognize the SKU.

Didn't upgrade/downgrade the bios.
I will reseat the cpu as a last resort. But I'm not hopeful.

>Another possibility: I have encountered certain cases with the X10 series will play dead if it is not convinced there is an active CPU fan plugged in to either FAN1 or FANA. Could there be a disconnected fan.?

Interesting!? I can certainly try plugging in another fan and see what happens. Though I'm guessing that since there are no lights on, this probably won't make a difference.

 

[Chris Moore]

No it's not FreeNAS. I asked the question here because the forums are much more active than the opensolaris forums I used to frequent.
I worked for Sun Microsystems and started with the original Opensolaris and migrated to OmniOS CE over the years.

Welcome to the FreeNAS forum. Perhaps we can lure you into using FreeNAS.
Is the system you are having trouble with using OmniOS?
I think that is the same version of OpenZFS that FreeNAS is currently using.

I do use it and it's completely dead. There are no leds on anywhere on the motherboard or connectors on the back.

That is not good at all. Did you try completely disconnecting power and then reconnecting it? That will sometimes get a system to come back.

Maybe I'll try this as a last resort, but I'm thinking a dead motherboard is most likely the culprit.

I think we are at that point now. If re-seating every connector, including the CPU, does not revive it, Supermicro has been known to take a board in for work even after the warranty period. You might want to try sending them a message.

 

[Ericloewe]

It's worth physically inspecting the board for damage, like blown caps, burnt traces, etc. Not that it's likely to result in an easy fix, even if you identify something, but it helps make sure the board wasn't destroyed by external forces.