Digital Ocean, Debian 12 64 and FreePBX 17 script

Hi,
I created a new Digital Ocean droplet with Debian 12 64, ran apt-get update && apt-get upgrade and then the freepbx 17 script.
All worked great, I can register the PBX, create certificates, change ports etc etc. However 2 things that dont work, and it may be Digital Ocean thats the issue.

System Admin - Network Settings fails to load

I can tweak that by changing the permissions of the /etc/network/interfaces file adding “other” to have read and write. Is there a better way to do this?

The only other issue I can see so far is I cannot change the eth0 or eth1 zone to any other type, specifically Internet. I have the Digital Ocean firewall configured, but I would prefer to use both FreePBX and DigitalOcean. Would there be any guidance on what to look at to change the eth0 zone to internet?

I have tried 3 installs so far and all have the zone issue.
I have tried changing to the Edge Modules with no difference.

Once i get the firewall sorted I can move on to more testing.

Thanks
Dave

Hi,

Same firewall issue here.

To have a working firewall a switch to network config via files in /etc/network/interfaces.d seems mandatory.

Other configs like netplan or systemd-networkd are not supported to secure the server.
Quite a limitation …

Have a look here in the code : firewall/Network.class.php at release/17.0 · FreePBX/firewall · GitHub

Thanks, looks like its the fix, I haven’t had a chance to test as I have had to jump onto a different project but will update the ticket when I do,
Best wishes
Dave

To anyone else finding this thread because of problems with a 2+ minute delay on boot with a cloud provider like DigitalOcean that uses systemd-networkd and cloud-init netplan to configure the network interface.

If you plan to continue to use systemd-networkd with any cloud provider that uses /etc/netplan/*.yaml files, make sure the ifupdown package is not installed.

dpkg -l | grep ifupdown
apt purge ifupdown

Otherwise it causes a 2+ minute boot delay because

systemd-networkd-wait-online.service

times out. The reason is because the ifupdown package causes cloud-init to switch from using netplan network renderer to using eni renderer.

https://cloudinit.readthedocs.io/en/latest/reference/network-config.html#network-output-policy

After purging ifupdown you need to regenerate the /etc/netplan/*.yaml file

rm /var/lib/cloud/data/instance-id
cloud-init init --local

Now test it

netplan try --timeout 10

If you don’t lose network connectivity hit enter to apply the configuration. Reboot should no longer have a 2+ minute delay.

I think I may have installed the ifupdown package when doing something else not knowing it would cause network problems on reboot after cloud-init clean