BIOS Update prevents DSM from booting

[valix]

Hi there,

I updated the BIOS to the latest version, and since then, DSM doesn't boot anymore—only the bootloader does. The setup had been working flawlessly since February with the same hardware and DSM 7.2.1U2 U3 and U4 (after I initially got it running).
 

My build:

• AMD Ryzen 5 Pro 5650G
• Gigabyte MC12-LE0 BIOS F13 -> Updated to F15
• 2x 16 GB ECC RAM
• 2x 1 TB WD Red NVMe via Bifurcation adapter (PCIe x16 Slot)
• 2x Intel i210 GbE (onboard, not in use/disconnected)
• ASUS XG-C100C, older revision with Aquantia AQC107 (PCIe x4 Slot)
• Aspeed 2500 BMC
• 3x MG10 Toshiba disks
• Loader: RR 24.1.3 SA6400 and later RR 24.8.6, with "all" modules
• DSM: 7.2.1 u4

 

The reason I updated the BIOS was that the previous version couldn't let my APU run at TDP 35/45 Watts, despite the option being offered. Even manual TPP, TDC, and EDC settings were ignored.

When I realized that I couldn’t get the NAS to boot after the update, I decided to update the loader first to the latest RR version (24.8.6) before attempting a rebuild. I burned the image onto the stick and tried to set up the new bootloader.

However, I still have the issue that the loader doesn't boot DSM. One of the main changes in the BIOS was an AGESA update to AGESA 1.2.0.Ca. Do you think the AGESA update might be influencing this issue? However, I previously updated from F11 to F13 in March without any issues, and that also included an AGESA update. Since I did not expect such issues, I did not unmount the NVMe Read-Write Cache in advance. Might this be a fault?
 

What I have already done:

• Restored the BIOS defaults and used only the most important settings, like bifurcation for the M.2 adapter for the NVMe cache.
• Set the boot mode to UEFI.
• Rebuilt the loader as SA6400 (the only device with the latest Zen architecture support (Zen2), with all modules), as before.
• Used the option in the loader to search for an existing DSM installation, which it found. (Was this a good or bad idea?)
• Booting led only to a pingable IP but no reachable web front end for "migration".
• Tried other loaders for AMD V1000 and AMD R1000, but all led to reboot loops while trying to boot DSM. I will not try any Intel based DSM Version since I realized that it didn't work neither on older Carizzo nor on the newer Zen3. BTW Carrizo runs flawlessly as DS923+ (AMD R1000 Architecture)

 

What I will try next:

Clear the CMOS settings and try the SA6400 loader again with the minimum BIOS settings. ->Successfull? At least I can start the migration.

 

Questions:

• Does anyone know a way to reset the loader stick to default without reburning the image repeatedly? I thought it would be sufficient to delete the user-profile.yml (or whatever it’s called), but that doesn't seem to be enough.
• Does anyone have any ideas?
• Is there a BIOS option that I should check?

 

I am running out of ideas.

 

 

BTW: An ActiveBackupForBusiness backup from the day before is stored on the backup NAS. Data loss wouldn't be ideal, but it might be acceptable if absolutely necessary.
BTW2:This post is intended to help solve my problem and to serve as documentation for anyone who might encounter the same issue in the future.

 

Edit: First it seemed that it worked, I could access the Recovery Webfrontend, but the revovery failed since it weren't able to execute the announced reboot. So I rebooted it manually, after I waited over 15 minuts. And now it doesn't bot anymore DSM.

Below the screenhots of the last try after CMOS Clear:

 

 

 

 

Then it started to the recovery web page and after a manually reboot, because the self initiated reboot didn't worked, I got this and an unpingable device:

 

Then I rebooted to "Recovery" in GRUB menu.

 

And now it looks better:

Edited August 31 by valix

[valix]

But still somehow the same, no reboot after reboot anouncment (Ping would have been unsuccessful while rebooting). Therefore I rebooted the machine again manually.

 

 

But then it remained unreachable:

Edited August 31 by valix

[cferra]

Hello there, things that I would suggest to start:

 

1. Enable hotplug for your drives - should be under sata settings

2. Disable "Above 4G decoding"

3. Disable Resizable Bar

4. Disable all C-States

5. Disable TPM

 

Try that and see if you can get it to boot, if it does, you can start to re-enable features until you figure out what causes the system to hang when booting.

[cferra]

Does anyone know a way to reset the loader stick to default without reburning the image repeatedly?

 

rm -f /mnt/p1/user-config.yml /mnt/p2/* /mnt/p3/users/* /mnt/p3/users/*-dsm

Edited August 31 by cferra

[valix]

Hi, thank you for sharing your ideas.

There is no hotplug option in BIOS.
"Above 4G decoding" -> was already disabled

"Resizable Bar" -> was already disabled

"C-States" -> now disabled

"TPM" -> now disabled

No change. At 58% installation progress it jumps to the reboot announcement without rebooting, when I reboot manually it remains offline after the RR loads the DSM.

I alos tried 7.2.2 without getting any further.

[valix]

1 hour ago, cferra said:

/mnt/p3/users/* /mnt/p3/users/*-dsm

The folder /mnt/p3/users/ doesn't exist for me.

[wjz304]

rm -f /mnt/p1/user-config.yml /mnt/p2/* /mnt/p3/users/* /mnt/p3/*-dsm

Ignore files that don't exist

 

and try "Advanced menu -> Switch direct boot: true"

[valix]

@wjz304 Thank you for the clearification of resetting the stick config.

Meanwhile I removed all prod disks and inserted for testing a brand new single disk.
 

I tried the direct boot, but the result is the same:
After the first boot of DSM where I install / load the PAT file, an expect the first reboot, nothing happens, except the countdown of the 10 minutes despite the system remains up and running......

 

[wjz304]

35 minutes ago, valix said:

@wjz304 Thank you for the clearification of resetting the stick config.

Meanwhile I removed all prod disks and inserted for testing a brand new single disk.
 

I tried the direct boot, but the result is the same:
After the first boot of DSM where I install / load the PAT file, an expect the first reboot, nothing happens, except the countdown of the 10 minutes despite the system remains up and running......

 

Try to remove the static ip settings.

[valix]

The last try was already without static ip. The last try from 1 hour ago was already with DHCP as IP.
192.168.3.4 -> static
 192.168.3.2x - .250 -> dhcp.

 

This is the current and since then unchanged state, I  posted one hour ago. Prior, I think I missed to list all information.

•     brand new unformatted disk - removed all former disks
•     reflashed the USB stick (was before you confirmed that the command was right for resetting the stick)
•     Build a new loader with direct boot and no static ip, the rest remained like in my first post
•     I got to the Install WebPage of DSM (also the Syno Assistant found it) and chose the DSM_SA6400_69057.pat as DSM file. The the progress percentage jumped from 0 to 8 % to 40, 56%, 57%, 58% and said then "rebooting", and since then the 10 minutes refresh counter loops and the device never startet the shutdown part of the reboot, since it was permanently pingable, proper reboot would cause an unresponsive ping.
• If I now restart the device manually via my IPMI access, it gets stuck in the DSM loading process without a reachable IP. The IP will only be accessible for a short time as long as the loader is active

Edited August 31 by valix

[valix]

Interestingly, I just downgraded to F13 and the problems persits: No auto reboot after DSM Installation, I assume that the PAT-installation fails somehow. (Current loader build for DSM 7.2.2) . Of course the BIOS loaded its defaults... but I ran it at almost default settings before (except bifurcation).

I will try with a loader for 7.2.1 and I will try the older BIOS with RR 24.1.3 which I used before.

 

 

Edited September 1 by valix

[valix]

All the same.... I m thinking about going for virtualization....

[valix]

Well, now I installed proxmox and setup a VM with AMD Rome Architecture CPU settings and SA6400 bootloader and it works fine. I was able to mount the physical HDDs and the NVMes to the NAS. The BIOS is F15.

And all looks like before: no data loss, same NVMe Cache Array....

[valix]

After realizing that I was in trouble if I wanted to expand my storage through a disk swap, I went from virtualization back to bare metal and tried the Auxxilium Loader which I set up also as SA6400, which worked fine on the first try and also with the latest BIOS (F15), and I got my setup working again.

Edited September 8 by valix