I know this is more Linux related but the problem is with the standard FreePBX distro, CentOS based. Internet access is via a router and dynamically assigned IP via DHCP (but always the same address). Problem happens after unexpected power outage, when FreePBX boots and internet is not available yet, neither the DHCP I think (router may boot slower). In such case, FreePBX never again tries to obtain the IP and as a result, server is ‘dead’. It is enough to reboot it once the router us up again, but that requires a physical access. Is there a way to fix it in a nice way, not putting systemctl call in the cron or anything like that?
I often use static DHCP on devices, so for testing or troubleshooting I can plug them into another network and avoid IP address conflict or unreachability. However, that’s not very useful for a PBX (the extensions won’t work and if using IP auth, the trunks also won’t work), so maybe you should just configure the network interface statically and call it a day.
If your power outages are brief, consider putting network gear and the PBX on a UPS; in addition to avoiding this issue, these systems will be less likely to suffer corruption from outages.
If your power outages are lengthy (and a generator is not in your budget or otherwise impractical), consider putting the PBX in the cloud. It will function normally during an outage and calls can be made and taken on mobile phones.
I second @Stewart1 . There is zero reason to have a production PBX on DHCP. Needs to be a static IP configuration to avoid issues like this.