SOLVED Supermicro A2SDI-8C+-HLN4F-O died ...

[Patrick M. Hausen]

Hi all,

this weekend my home server simply ceased working. The IPMI event log shows:

Code:

4,2018/07/07 12:28:24, #0xff,Processor,IERR - Assertion
5,2018/07/07 12:30:19, #0xca,Watchdog 2,Timer Interrupt - Assertion
6,2018/07/07 12:30:20, #0xca,Watchdog 2,Hard Reset - Assertion

Now the system won't boot at all.

Any ideas? Is this a "C2000" CPU and possibly prone to this problem: https://forums.freenas.org/index.ph...s-system-fault-may-lead-to-dead-system.50314/ ?

Thanks for any hints,
Patrick

 

[MrToddsFriends]

Is this a "C2000" CPU and possibly prone to this problem: https://forums.freenas.org/index.ph...s-system-fault-may-lead-to-dead-system.50314/ ?

No, the C3758 on your board belongs to the C3000 (a/k/a Denverton) series.

 

[Patrick M. Hausen]

I opened a support case with my supplier - I'll report how that turns out.

Patrick

 

[Patrick M. Hausen]

They are RMAing the board ...

Kind regards
Patrick

 

[tedaz]

They are RMAing the board ...

Kind regards
Patrick

It looks like a hardware failure, RMA is necessary.

Sent from my SM-G900F using Tapatalk

 

[Patrick M. Hausen]

They don't have the same board available and I have to return it for repair by Supermicro - takes a couple of weeks. So they offered me a X10SDV-4C-TLN4F-B instead - looks good to me. Half the cores but Xeon D instead of Atom ... should pack a punch ;)

Patrick

 

[Patrick M. Hausen]

Turns out it wasn't the mainboard or the CPU but the memory. Because the replacement board they sent me showed the exact same symptoms.

After swapping the power supply which did not help I removed the DIMMs and tried them two at a time. Looks like at least one is broken.

What puzzles me is that the system simply refused to boot with the module installed without a clear indication of what was wrong.

I mean, of course even ECC modules can break for various reasons and they do. But all the servers we run in the data centre of course scream bloody murder, kernel panic and reset first - but then simply deactivate a broken DIMM and reboot with less memory so service is restored.

And this Supermicro system although tiny is supposedly "server grade" hardware and costs quite a bit more than your average desktop PC ...

I'll have to talk to the supplier tomorrow.

Kind regards
Patrick

 

[BigDave]

The majority of boards manufactured today, no longer seem to come with an onboard
system speaker any more, there are pin blocks provided to hookup add-ons, but "on-board" seems
to be a thing of the past.
I'm wondering if your RAM issue might have been brought to attention if you had a working speaker?