Any tips or tricks to get a local controller to manage switches and APs at a remote site over a VPN (not using Route10) ?
When you set up the controller, you were provided 2 DDNS hostnames; one beginning with âlocal.â
Provided the subnet for the IP that the hostname beginning with âlocal.â is reachable over the VPN tunnel (and TCP port 443 isnât blocked in the process), you should be able to provide that hostname to Alta devices and theyâll communicate over the tunnel.
I would expect it to be that simple but it has not been my experience so far. I enter the local. address in the AP however it never connects to the on-prem controller at the remote site. All traffic is allowed and DNS resolves correctly, however it never connects and has to be factory reset to re-join the cloud controller.
From that side of the link, can you access the controller with the hostname that starts with âlocal.â from a computer?
From the AP I can ping the local. address and hit the controllerâs page with wget
Did you use http://local. or https://local.?
https
Can you DM me the link that youâre pasting into the device? Emphasis on DM for security purposes.
That URL is good, thatâs really weird given youâre able to curl the URL from the AP. What VPN protocol is in use? Maybe itâs MTU related or something like that.
The VPN tunnel is running wireguard via two Firewalla devices in a site-to-site configuration
Rapidly running out of ideas. What are the 2 subnets in play at both sides?
Subnets are different and non-overlapping /24âs on each side
Well, in that case, itâs time to call in the big guns. @Alta-Jeff do you have any insights that might explain why an Alta device wonât talk to a local controller over a VPN tunnel despite being accessible by all accounts, specifically WireGuard S2S (non-Alta if it matters)?
Iâd have to look at device and controller logs, to know whatâs going on, but the first thing I would do is ssh to the device and see if you can hit https://localâŚddns.manage.alta.inc/api/status (with and without the local prefix).
root@<removed>:~# wget https://local.<removed>.ddns.manage.alta.inc/api/status
Downloading 'https://local.<removed>.ddns.manage.alta.inc/api/status'
Connecting to 192.168.<removed>:443
Writing to 'status'
status 100% |*******************************| 11 0:00:00 ETA
Download completed (11 bytes)
root@<removed>:~# vi status
{"ok":true}
Looks like youâre able to reach via local, at least. Can you try this on the AP/switch/router:
- /etc/init.d/rc stop
- echo https://.ddns.manage.alta.inc >/cfg/domain.txt (yes, without the local; rc will automatically detect whether the local prefix is necessary)
- rm /cfg/rcapi.txt
- /etc/init.d/rc start
The AP/switch/router should then be able to be controlled by that controller.
Performed those steps on the AP. Now the device is showing âdisconnectedâ on the Cloud Controller and is not showing up as a device to add on the Local Controller.
Browsing to itâs IP shows this message:
Can you share the controller logs by running journalctl on the controller as root? Would be best in a DM, if possible (or gdrive link, etc.)