I’m running Truenas Scale on my homemade NAS and I want to upgrade my network to 10Gps.
I bought a Dell X710-DA2 (used one) on ebay and I’m facing a problem.
When I boot my NAS with the card installed, it fails at the POST (so even before BIOS loading) with the following error :
0x06=mrcReadMPRErr
I’ve not been able to find any help on the web. The few other cases with this error message were linked to new RAM addition, not to a network PCI card.
When I remove the card, the system boots correctly and works as a charm.
Regarding the card, I’ve plugged it in my PC to upgrade the firmware and I had no issue.
It looks like a compatibility issue, but it sounds weird for this kind of hardware.
Here is my config :
Motherboard : Supermicro X11SSL-F
CPU : Intel Xeon E3 1245 v5
RAM : 4x16GB ECC
Boot disk : 2x128GB SDD (mirrored)
VM and app disk : 2x600GB SSD (mirrored)
Data disk : 4x2TB (RaidZ1)
Can someone help me ?
Thanks in advance for your help
I’ve tried to disable the OPROM and even tried EFI option, with no result. At boot, the system beeps and displays the same error message.
This time I’ve captured the error message displayed very quickly just before : 0x18=mrcColdBootRequired
I don’t know if it can help.
Thanks
According to what i could find on the net, this error also point to a RAM error (during training)… at this point i don’t think that a cold boot will help, but try cost nothing (just power off - unplug or turn off the PSU switch, and keep the power button pressed for some seconds).
In your place i would also try to unplug all the RAM sticks and try just with one (check the manual for the correct slot to populate with 1 only stick).
In some thread of the old forum, people with similar problem have disabled the SMBus via jumper on the mainboard
Yes I saw these posts related to a RAM issue, but I don’t think it is the same case as mine. In these posts, they were facing the issue when adding RAM, which is not my case.
Moreover, my Truenas instance is working perfectly since years with its 64GB of ECC RAM. It never raised any issue.
I’ve opened a ticket to Supermicro in parallel. Maybe they will be able to find the root cause.
I’ll keep you posted in any case.
I’m still in contact with Supermicro support. For the moment we are just reviewing all the bios and firmware versions. My motherboard bios was not up-to-date, so I made the upgrade. I checked again the firmware of my NIC (Dell version) and I’m already running the last version available (9.64).
So far, I still have the same issue. I’m waiting for the support to come back with new ideas.
Following the Supermicro support advice, I’ve tried first to enable the “Above 4G Decode” option in the Bios, without any positive effect.
I’ve then tried to boot with only 1 RAM module and that time I was able to boot successfully ! I’ve also tried with 2 modules and it was again working. I’ve tried with 3 modules and even if the systems boots, IPMI and Truenas only see 2 modules (2x16GB).
I’ve swapped some RAM modules, suspecting a faulty one but I had no change.
I keep my config like that for the moment, waiting for the feedback from the support.
Thanks for reading this novel, and see you for the next episode .
IMHO i don’t think that the RAM are failing or similar: seems just that the mainboard can’t handle everything somehow in your config, before posting.
Have you tried to disable the SMBus jumper? As you see, tryng with less sticks have had good effect
I agree with your analysis. I think, for whatever reason, the system is not able to handle 4 sticks of RAM and the NIC at the same time.
My SMBus jumbers are enabled. I’ll try to disable when I’ve time.
I’ll tell you the result
The solution was the SMBus jumpers Ji2C1 and Ji2C2. When set to pin2 and pin3, the system boots with 4 RAM sticks and the NIC installed.
It seems these jumpers affect the addressing process of i2C.
To support more devices, they have to be disabled.
.
