SG-3100 Bricked after update, help :)

eyes

Hello All

I decided to update my Netgate 3100 to the newest version 23.05 from 23.01

I have had some issues with the online update in the past, (it got stuck during process) so i asked support for the firmware file to do it manuel via usb.

I followed the guide https://docs.netgate.com/pfsense/en/latest/solutions/sg-3100/reinstall-pfsense.html and patched via USB

After the patch was done i powercycled the device.

The device is now stuck in u-boot and just reboots endless. I have tried to install both on the eMMC and a M.2 drive.

Anyone know a way to get the default factory settings for U-boot and the environment variables. Mabye delete the old and paste in the defaults.

I think mine are wrong.

Best Regards
Simon

eyes

I managed to get it to work with the new firmware version. 23.05

First copy your current env settings with.

output of the 'printenv' command from the Marvell>> prompt

Then i deleted all in the environment of U-boot

Then i reset the environment to defaults.

After that i entered the values support emailed me.

setenv emmcboot 'fatload mmc 0:1 200000 ubldr.bin; go 200000;'

setenv factory 'setenv autoload yes; bootp; go $loadaddr;'

setenv m2boot 'scsi reset; fatload scsi 0:$m2dev 200000 ubldr.bin; go 200000;'

setenv recovery 'usb start; fatload usb 0:1 200000 ubldr.bin; go 200000;'

setenv setled 'i2c mw 67 ff 0; i2c mw 67 0 1; i2c mw 67 f 35; i2c mw 67 3 40; i2c mw 67 26 0; i2c mw 67 10 0'

setenv m2dev 1

setenv fdtfile armada-385-80300-0148-G00-X100.dtb

setenv bootcmd 'run setLED; run emmcboot;'

setenv boardrev R306;

setenv boardpn 80500-0148-G00

The additional values rely on the WAN MAC and serial of the device.

setenv boardsn your sn
setenv ethaddr 'your mac'
setenv eth1addr 'your mac'
setenv eth2addr 'your mac'

All these 4 values should be listed in the old env sittings you backed up.

When you entered all lines finish with saveenv to save the settings.

After that enter reset and boot the device.

The device will now boot into the onboard chip and work with flashing reinstall

marvell >> usb reset

followed up by

marvell>> run recovery

FSC830

Just reading this solution, but to be honest: without the issue its hard to understand when this should be used.

I am running a SG-3100 too and do not have had any issues when updating from 22.05 to 23.01.

So my guess is you did an update to 23.05 and run into an issue?

My recommendation: please do not remove the issue when posting a solution. So someone can compare, if the solution (better the issue) matches for him.

Regards

eyes

@FSC830

Sorry FSC830

I dunno why it was flagged as deleted. Its there now.

My device got stuck in the u-boot bios after trying to update to 23.05. And nothing happened after doing run recovery.

It just kept rebooting, that was the issue.

Best Regards
Simon

FSC830

Thanks. Yes, it was marked as deleted - for what reason ever...

I did update my SG-3100 today (online update), but as always I requested an recovery image in advance.

The update was very smooth and without any issues.
In the next days I will update the second SG-3100 in the same way.

Regards

eyes

@FSC830

BTW. Did you upgrade your device with the M.2 storage?

If yes, did you enter any commands to get it to work?

Best Regards
Simon

FSC830

@eyes : I installed the M.2 devices abt. 1 year ago.
Actually I cant remember exactly if I did modify an boot option.
As far as I remember I started a new, clean installation with the recovery image and restored the configuration from a backup.
All was smooth at both devices.
The update yesterday did not need any modifications, just updating and rebooting and all was fine.
Only steps needed is to copy some modified pages for captive portal back to /usr/local/www.

Regards

stephenw10

Yes, the only time you should have to do anything with the uboot envs is if you change the boot device. So usually only if you add an m.2 SSD.

See step 7 here: https://docs.netgate.com/pfsense/en/latest/solutions/sg-3100/m-2-sata-installation.html

Hard to know exactly what happened to @eyes here without seeing the boot loop output. We have seen the uboot envs become corrupted in the past so that is possible. That would require rewriting them.

Steve