Search results

As requested. A little update: I was able to get opnsense installed on the eMMC. That went without a hitch by following the documentation. I then switched to proxmox, which I'm more familiar with. Proxmox does not allow installing on eMMC as a default option. You have to install in debug mode...

I'll get a photo next week as I left it at work. It does have the wifi mini PCIe slot but the second M.2 key is unpopulated. Since the BIOS update was looking for two HDDs I was a little concerned it wasn't going to sort out that the second slot didn't physically exist. I did notice the SIM card...

You're right, it was a nut on the power supply, not wifi. I'm so used to the only threaded port on an IO shield being wifi haha. As for the boot issues, it was a indeed a corrupt BIOS. I was able to use a CH341a chip programmer to update the top BIOS chip with the VEP1400-X-BIOS-3.48.0.9-23.bin...

Turns out it was an extreme friction fit between the screw posts and screw holes in the PCB. Also, on my unit the wifi antenna is mounted to the rear IO shield which holds it in place. Once removed the tolerances loosened up a bit. I just needed to brute force it a little bit more. I repasted...

Yep, I'm the one that bought that unit lol. It did come with brackets and as far as I can tell is NIB as advertised. As for rebooting, I think both BIOS slots are toast as I get a "ERROR: recovery capsule not found" as shown below. I haven't been able to get into the BIOS using del/F2 and have...

I'm going to have to deep dive this entire thread but just adding my story of getting sniped by this device. I got a NIB Vectra S1 branded version of the the 1445 with 32gb of RAM and a 960gb SSD. It wouldn't boot into an OS and I had no interest in Vectra software so I got DiagOS installed...

I have one of these with a pair of 14tb WD drives and it's been amazing. I haven't done any mods but keep thinking about it. Recently upgraded to 64gb of mem but even with 32gb it ran proxmox great with truenas scale and core virtualized. I don't run a ton of services but jellyfin and pihole...

This saved my bacon today. I was able to use SMCIPMITool from SuperMicro to remotely manage this. Used the following command SMCIPMITool.exe <IP Address> <username> <password> ipmi lan dhcp enable which oddly returns "static ip enabled" but login worked without hanging. The BMC reported that...

I wouldn't say that as myself and others have had significant trouble or been unable to even post using LRDIMMs. I do have a promising update though. I was finally able to POST my LRDIMMs but it's still a work in progress. I had to first boot with an 8GB RDIMM. From there I have to manually set...

I'm using m386a8k40cm2-cvfby which is 2933 which isn't posting. I may try to manually set memory timings to massage it past an initial memory training. On another thread others had inconsistent lrdimm compatibility and the eBay seller says 2Rx4 is best supported. I really don't want to have to...

Thank you both, that gives me the confirmation I need. I don't know if this helps but the main difference between the system logs that I found was: " spx_restservice: spx_restservice - - [2159 : 2159 WARNING][fru.c:135]Invalid FRU Header Version for ID : 6 - " For IDs 6-20. It only happens...

Hi Tugm4470, I also ordered a MZ32-AR0 (rev1.0) with an Epyc 7d12 from you. The board and CPU are good with normal ECC memory. I'm trying to use 64gb LRDIMMs from Samsung (M386A8K40CM2-CVFBY) which are on the QVL for the motherboard. Have you tested if the 7D12 is compatible with LRDIMMs?

Along these lines CDGH work on my MZ32 for regular ECC DIMM but for the life of me I can't get an LRDIMM to post. Has anyone tried LRDIMMs with any success?

I'm having the same issues on my MZ32. In BMC is shows 2x the total memory, and the DIMMs aren't the ones populated (A/E channels when I only have one dimm in C channel). The BIOS also reports 2x total memory but under the memory settings/info page it accurately shows the single DIMM in C1...

I was able to update my BIOS to the stock rev3.0 (M18/R34) that's current on the GB site without issue. The most recent R28 for rev1 gave me an infinite power cycle loop, but the R25 worked too (my board came with R19). I am using a 7D12 CPU though, so the R28 boot cycle could be a result of the...

Ah, yeah I garbled that since I was focusing on NICs per OP's original question. I forgot only the 6036 *G* had the VPI out of the box in contrast to how their VPI marketed NICs are mode switchable. Their marketing about VPI vs Gateway makes it seem like pure eth with a mode switch was possible...

That's exactly what I've been saying from the start...

Did this work out for you? Seems like just shorting the JP2/FNP would suffice as the cards don't seem to be never burned, just wiped from what they had before.

Explain? Like I said above, the rebrands introduce licensing issues (and the rebrands are the majority of what you find cheap). Mellanox firmware itself hasn't required a separate license for ages, but flashing a switch from rebrand to MLNX-OS is "here be dragons" territory. Ironically though...

This seems to be an issue with people not paying close attention to which SKU they buy. As far as I know, all 40gb NICs, there are 3 versions of ConnectX based products, pure ethernet, pure IB, and VPI which supports both but must be manually configured and rebooted to switch between the two (by...