SG-1100 Update gone awry

droppointalpha

Good day.

My father recently ran updates on his SG-1100. However, from his description the system never rebooted/recovered.
He attempted a power cycle but the system did not reboot.
He then attempted the factory reset button but this was also unsuccessful.

When he reached out to me, we walked through some troubleshooting before I ultimately talked him through making a console connection via USB.

Unfortunately, all I have is pictures but I'll transcribe out the highlights as best I can.

First up - versions
U-Boot version is 2017.03-armada-17.10.2-g7a6581a-dirty
FreeBSD/arm64 EFI loader, Revision 1.1
Unfortunately, I don't see anything to indicate the version of pfsense his machine is attempting to boot into.

Now, the console transcribe (please forgive me any minor typos); I will attempt to visually replicate as I see it in images.

FreeBSD/arm64 EFI loader, Revision 1.1
(Fri Feb 10 20:26:39 UTC 2023 root@freebsd)
mmc_init: -95, time 23
   Command line arguments: loader.efi
   Image base: 0x5000000): FDT_ERR_BADMAGIC
   EFI version: 2.05s configured. Please configure
   EFI Firmware: Das U-boot (rev 0.00)>" command.
   Console: efi, comconsole (0)
   Load Path: bootefi configured. Please configure
   Load Device: bootefit addr <address>" command.
ERROR: cannot open /boot/lua/loader.lua: invalid argument.
## Starting EFI application at 05000000 ...
WARNING: Invalid device tree, expect boot to fail
Type '?' for a list of commands, 'help' for more detailed help.
OK Application terminated, r = -2
reading efi/boot/bootaa64.efi
839196 bytes read in 36 ms (22.2 MiB/s)
reading armada-3720-sg1100.dtb
17970 bytes read in 12 ms (1.4 MiB/s)
## Starting EFI application at 05000000 ...
Card did not respond to voltage select!
mmc_init: -95, time 23
Scanning disk sdhci@d8000.blk...
Found 4 disks

And that is where the image cuts off.

I will have him kick off a TAC request but I have also cruised here looking for others with similar circumstances and trying to work additional angles.
Unfortunately, while I have an SG-1100 as well, I'm not as familiar with the workings of these.
Fortunately, we have a configuration saved!

So - is there a recovery partition we can use the console to restore from?
Is a complete reinstall the only option?
Are there diagnostic commands I can have him run?
What other posts that I might have missed can I read.

SteveITS

@droppointalpha Reinstalling is less complicated than it sounds especially if you have console already, and is very likely the fastest way to a known good system. TAC will point you to https://docs.netgate.com/pfsense/en/latest/solutions/sg-1100/reinstall-pfsense.html

Just need to burn the USB stick.

stephenw10

@droppointalpha said in SG-1100 Update gone awry:

ERROR: cannot open /boot/lua/loader.lua

That's a boot loader error. Do you know what version it was updating from?
There was a known bug with that a few versions back.

Re-installing is almost certainly the quickest way back.

Steve

droppointalpha

I appreciate the replies and I figured a reinstall would be the recommended path forward considering it wasn't getting past boot.

But for my personal edification, is there no read-only section of storage with a reserve or roll-back copy of the OS and/or boot loader?
Or is this where, if once desires some durability, then go ZFS?
Answers need not be for the SG-1100 in particular - I'm going to be expanding my network to accommodate two additional lines in and those will be served by larger appliances.

Also - what does it mean at the end that it found 4 disks? What is the usual partition setup?

@stephenw10 I figured as much; I've borked a boot loader or two before (but none of those were U-boot).
Is this case, I'm not sure he was several versions behind. Mental count of when I helped him setup the device and the updates I've run since then, its possible but that would mean he's never updated ever. He's not a big slacker when it comes to updates and he has in/out of the system learning and experimenting since I set it up.
Unfortunately, the only version information I have is what I transcribed from the images of console output. So I have no way to know for sure.

stephenw10

There's no recovery partition in the default install.

Yes, use ZFS for filesystem resiliency.

That 4 disks line is somewhat confusing. It's output from uboot and actually shows partitions or bootable system devices. 4 is the expected number.

droppointalpha

@stephenw10
What are the four partitions?
boot, main os, swap, and ???

stephenw10

There is an EFI partition also. But there are also 4 geom devices:

23.09-DEVELOPMENT][root@1100-3.stevew.lan]/root: geom disk list
Geom name: flash/spi0
Providers:
1. Name: flash/spi0
   Mediasize: 4194304 (4.0M)
   Sectorsize: 512
   Stripesize: 4096
   Stripeoffset: 0
   Mode: r0w0e0
   descr: mx25u3235f
   ident: (null)
   rotationrate: unknown
   fwsectors: 0
   fwheads: 0

Geom name: mmcsd0
Providers:
1. Name: mmcsd0
   Mediasize: 7820083200 (7.3G)
   Sectorsize: 512
   Stripesize: 512
   Stripeoffset: 0
   Mode: r1w1e3
   descr: MMCHC DG4008 0.1 SN 2EC51400 MFG 06/2017 by 69 0x0000
   ident: 2EC51400
   rotationrate: 0
   fwsectors: 0
   fwheads: 0

Geom name: mmcsd0boot0
Providers:
1. Name: mmcsd0boot0
   Mediasize: 4194304 (4.0M)
   Sectorsize: 512
   Stripesize: 512
   Stripeoffset: 0
   Mode: r0w0e0
   descr: MMCHC DG4008 0.1 SN 2EC51400 MFG 06/2017 by 69 0x0000
   ident: 2EC51400
   rotationrate: 0
   fwsectors: 0
   fwheads: 0

Geom name: mmcsd0boot1
Providers:
1. Name: mmcsd0boot1
   Mediasize: 4194304 (4.0M)
   Sectorsize: 512
   Stripesize: 512
   Stripeoffset: 0
   Mode: r0w0e0
   descr: MMCHC DG4008 0.1 SN 2EC51400 MFG 06/2017 by 69 0x0000
   ident: 2EC51400
   rotationrate: 0
   fwsectors: 0
   fwheads: 0

I'm not entirely sure what uboot is looking for that value. Let me see.....

stephenw10

Nope, still unsure!