That sounds like a “somehow crashed” BMC. The only thing that seems to work then is to unplug the power supply, physically remove the CMOS battery, wait for while (eat lunch etc.) and then reseat/reconnect everything. Just using the CMOS clear jumper doesn’t seem to be enough.
Then reflash the BMC and BIOS with the same versions via IPMI and do not (!) keep the existing configurations. After that it should post again with graphical output.
It could also help to use the KVM graphical output and not a physically connected display until it behaves “normally” again consistently.
Good luck with ASRock Rack’s support, as mentioned they don’t even respond to me anymore.
(edits due to mixed up languages - doing too many things simultaneously and began to type phonetically )
BMC was fine… I flashed it last night to 3.1 then 3.2 while I had my 200GE installed… it ran all night just fine.
This morning I put the 3700x in… no video/boot so went back to BIOS v3.10 and tried again… no video/boot.
I can try installing the 200GE again to double check I didnt brick anything, but with the 3700x installed, I am able to see sensors/logs, power it off & on in the IPMI.
Have you tried the suggested method with the removal of the CMOS battery and cutting power to the PSU?
Do you get a graphical outout on the “Remote Control” KVM after the battery removal stuff with a motherboard + CPU + RAM only after powering on the system (disconnect any physically connected display)?
Yes that sounds vaguely familiar. I think I experienced something like that when I switched from the 2600X I had on load for a while prior to the Ryzen 3000 launch the first time I tried to power on the motherboard with a 3700X.
My “gut feeling” suggests that a part of the BMC crashes somehow that requires a complete separation from any power source (PSU/CMOS battery).
I think another user experienced something similar, too.
Well, at least they seem to pay attention to this thread
(Cynicism) Can someone make screenshots of an UEFI editor and show how to click “File -> Import AMD binary/AGESA” and “Export as flashable file” so we can move a few AGESA versions up the queue?
You can boot a Fedora 30 Live system where you get to a nice desktop after a short while and during the boor sequence there are only nice green “OK” lines? (That’s the way it is working on my X570 Taichi with AGESA 1003 ABBA, on the X460D4U with AGESA 1003 x? I only get to the error messages (cp. screenshots some posts above)
What are your exact boot settings?
UEFI - CSM Disabled
UEFI - CSM Enabled
CSM Enabled - Legacy boot
?
Could you test if VMware ESXi 6.7 U3 boots on your motherboard unit without a “Purple screen of death”?
All three lead to the same behavior on the X470D4U/3700X (BIOS 3.20) and boot just fine on an ASRock X570 Taichi (BIOS 2.00 with 1003 ABB or 2.10 with 1003 ABBA).
So I’m about 99.9999 % sure that it is as his wendellness has stated in the linked thread that the AGESA is outdated meaning RDRAND is still broken leading to various OSes not working properly.
Again the only way we got is waiting for ASRock Rack to release a new BIOS, maybe if we’re not naughty we’ll get it for Christmas?