Hi,
Is it safe to do a „sysupgrade“ (without „-n“) with an image from a different domain (within the Neanderfunk area, for example „dias“ → „hlg“) to switch sites, or do you have to reconfigure the router completely, i.e. with a config reset?
I’m going to try adjusting the URL for the auto-updater now — that seems to be the usual way to do it after reading other threads about domain changes.
Yes, that works without any problems. Both the host key and the node name as well as the setting are retained, it just registers in a different domain then.
It switches the WiFi channel on the occasion and especially the identifier for the wifi-mesh to the setting of the new domain.
We have a helper script called „flash“. You can just throw the path of the sysupgrade directory (URL from the server) into it and the script will find the right image on its own and flash it.
To my shame, the flash command didn’t work because the „on-board
If the local flash helper aborts with rc=20, I would go the manual route only on the affected node and with power connected: load the sysupgrade image matching the target domain path to /tmp, check the file size or checksum, and then execute sysupgrade -c /tmp/<image>. Before that, back up the running configuration and the model name; using the wrong Gluon target would be more dangerous here than the actual domain switch.
Which two domains were yours (dias → hlg?) and which Gluon version or device target was running on the node? That way the cause of rc=20 can be narrowed down better.
If the target doesn’t match, fortunately the supergrade refuses it anyway. To override it there with -f, that’s something you should really think twice about.