I was raised with the rule that servers should always use static IP addresses. That still seems obviously necessary for foundational services such as domain controllers, DNS, DHCP, routers, and other infrastructure that must be reachable during an outage. But for ordinary application servers, is a manually configured static address still the best approach? Do modern applications and scripts generally use DNS names, or are hardcoded server IPs still common? Would DHCP reservations provide enough stability, or am I overthinking this?
4 Answers
There is no universal rule that every server must be manually static. The real questions are whether the system needs a stable identity and whether it must function when DHCP or DNS is unavailable. In small or less resilient environments, manually configured addresses reduce dependencies. In larger, automated environments, reservations, IPAM, DNS, and configuration management usually scale better. Just make sure the DHCP service is redundant and that critical recovery systems do not depend on it.
Applications and scripts should use DNS names or service discovery rather than hardcoded IP addresses whenever possible. If the name resolves correctly, the application usually does not care whether the address came from a manual configuration or a reservation. Hardcoded IPs create unnecessary work during subnet changes, disaster recovery, migrations, and failover.
You’re not nuts—the important thing is that critical systems have predictable addresses. That does not always require configuring the address manually on the server, though. DHCP reservations can provide a stable address while keeping address management centralized. I’d normally use true static addresses for DHCP and DNS servers, gateways, network appliances, and anything needed to restore basic connectivity. For less critical systems, reservations are usually a good compromise.
A practical policy is to divide systems by how essential they are. Keep manually configured addresses on core infrastructure and critical failover endpoints. Use DHCP reservations for application servers, printers, access points, cameras, and similar devices that need consistent addressing. Use ordinary DHCP for clients and disposable or cloud-based workloads. Document everything in IPAM instead of maintaining a spreadsheet.
This also makes disaster recovery easier. A server can receive a different address in a recovery network through a different DHCP scope, while clients continue connecting through the appropriate DNS name.
Reservations still depend on DHCP, so infrastructure needed to bring the network back should not rely on them. Redundant DHCP and monitoring make reservations much safer for ordinary servers.