i just had several storms that made the power flicker on and off several times last night.
Here’s the basic layout:
pfsense → arista switch → server rm switch → living room switch, then all other switches are downstream from there.
The ones specifically mentioned above are connecting to the controller, which is on a proxmox ct connected to the living room switch (used a tteck script that afaik pulls the official linux controller software). Also, these 2 unifi switches (server rm and living rm) were on UPS’s.
The downstream switches are not connecting to the controller. Also, there are 2 WAPs that are both downstream and use POE off the (offline) switches… I don’t think any of them are on UPS’s (though some may be… I have a lot of electronics and UPSs lol).
The switches and APs that are not connecting all show power right now, so it’s not like they are completely dead.
Any ideas on how to get these downstream switches to connect to the controller?
One other thing I should point out… the switches are on the default lan… 192.168.55.x whereas the controller is on the private vlan 192.168.70.x. This has never been a problem, and is not a problem for the (living rm) switch that I am connected to or the upstream unifi (server rm) switch. Only the downstream switches are affected.
I should have mentioned the downstream switches aren’t only not connecting to the controller… the downstream devices and ap’s are offline as far as device connectivity goes.
So, I got everything going more or less…
If you use the “reset” pin, be prepared to have to ssh into the switch to get it back on board. and if it’s changed ip, you have to delete it from your controller in order to readopt it. kiss your settings goodbye.
here’s the link to the ssh instructions:
yes, the title is “layer 3”, but the ssh portion applies to all unifi switches. You might need the user/pass from the controller’s “advanced” section, or the user/pass might be ui/ui. I had to use both.
Also, make sure if you have your controller on a different vlan from the switches (e.g. the switches are on default lan and controller is on private vlan in my case), make sure the switches can access 8080/tcp on the vlan where the controller is located before doing all this.
Bit of a mare then - reminder that there is a config backup facility in the controller, so assume once kit readopted restoring from backup should get things sorted?
never saw the email notification on this reply… must have gotten buried in other emails at the address I use for public accts…
Anyway, the short answer is yes, there is a utility to create backups of configs in the controller. The longer answer is the following:
I’m running the controller in a lxc put together by tteck that runs on proxmox. I believe it essentially repackages the official release into a lxc, but not sure.
for some reason backups weren’t working and I hadn’t checked since my recent move of the controller.
I’m not sure the backups would have helped, given the IP changes that also occurred. I’m just not sure.
I could have rolled back the lxc (I have backups of those) but I think the rolled back controller would have been looking for the old IPs at the very least.
