Failover vs High Availability: What Home Needs

Failover and high availability solve different problems, and confusing them is the most expensive planning mistake in home networking. Failover protects against your internet line going down by switching to a backup uplink. High availability protects against your firewall hardware dying by running a second identical firewall that takes over instantly. Most homes only ever need failover — in my years running this network the internet line has failed a handful of times a year, while the firewall hardware itself has never once been the cause of an outage.

I see people build the wrong one constantly. Someone terrified of internet outages spends a weekend configuring a two-firewall CARP cluster — which does absolutely nothing for an ISP outage, because both firewalls share the same dead uplink. Someone else runs a single firewall with no backup line and is surprised when a fiber cut takes the whole house offline for six hours. Each built for the wrong failure. This article draws the line clearly so you spend your redundancy budget on the failure you actually face.

If you have not built either yet, the multi-WAN and internet failover overview covers the failover side and the OPNsense setup guide walks the config. Here we decide which kind of redundancy your home actually needs.

What Is Failover, Really?

Failover is one firewall with two internet lines, switching to the backup when the primary dies. The single box stays in charge the whole time; only the uplink changes. When your fiber drops, the firewall detects it within seconds, reroutes traffic to your cellular or secondary line, and switches back when the primary recovers. It defends against the failure homes actually experience: the line, not the hardware.

This is what I run at home, and the reason is simple honesty about what breaks. In years of operating this network, the thing that goes down is always the internet connection — a fiber cut, an ISP maintenance window, a contractor’s backhoe two streets over. My fanless firewall itself has never been the cause of an outage. So my redundancy lives where the risk is: a second WAN, a cellular backup modem, and a gateway group that switches automatically. One box, two paths to the internet. That is failover, and for the overwhelming majority of homes it is the entire answer.

What Is High Availability and CARP?

High availability is two firewalls acting as one, so that if the active unit fails, the standby takes over in under a second. On OPNsense and pfSense this is built with CARP — the Common Address Redundancy Protocol — which gives the pair a shared virtual IP, plus state synchronization so existing connections survive the handover. It protects against the firewall itself failing: a dead power supply, a kernel panic, a botched update.

The mechanics are genuinely clever. Both firewalls run continuously; one is master and holds the shared virtual IP that your devices use as their gateway. A synchronization link keeps the standby’s configuration and connection state matched to the master, so when the master drops, the standby claims the virtual IP and traffic continues with barely a hiccup — often without even resetting active sessions. The pfSense high availability documentation and the OPNsense CARP how-to both walk the full setup. It is real, it works, and it is the right tool — for a problem most homes do not have.

What Is the Key Difference?

Failover protects against line failure; high availability protects against box failure. That single distinction decides everything. An HA pair sharing one internet connection does not survive an ISP outage, because both firewalls lose the same uplink at once. A single firewall with two WANs does not survive its own power supply dying. They guard opposite failure modes, and knowing which you fear tells you which to build.

Here is the trap in one sentence: high availability makes your firewall redundant, not your internet. People assume “high availability” means “always online,” and build a CARP cluster expecting it to save them from outages — then a fiber cut takes both firewalls offline simultaneously and they are baffled. The cluster did its job perfectly; its job was never the internet line. Conversely, failover makes your internet redundant, not your firewall. If your one box dies, two WANs do not help. The two are complementary, not interchangeable, and the whole point of choosing is to match your spending to the failure that is actually likely. Let me lay them side by side:

Why Is High Availability Usually Overkill at Home?

Because the failure it prevents is rare and the cost is high. A fanless firewall appliance with no moving parts is extremely reliable — the most likely reasons it goes down are a reboot for an update or a power blip, both of which are cheaper to address other ways. Doubling your hardware and roughly doubling your configuration complexity to survive a once-a-decade hardware death is a poor trade for most homes.

Think about what actually takes a home firewall offline. Updates — solved by scheduling them when nobody cares, not by a second box. Power outages — solved by a UPS, not by CARP, and if you already do UPS-backed power for the rack, your firewall rides through the same battery. Genuine hardware failure — real, but rare enough that a spare appliance on the shelf, ready to restore a config backup to, covers it for a fraction of the cost and complexity of a live HA pair. I keep a config backup and a serial console cable in the drawer for exactly that day. Spend your redundancy budget where the risk lives: on a second internet line, not a second firewall. That is not me being cheap — it is matching money to probability.

When Does High Availability Actually Earn Its Keep at Home?

When downtime during a firewall reboot is genuinely unacceptable — a real home office where a dropped client call costs you, a home lab hosting services other people depend on, or a setup where you cannot physically get to the rack quickly to swap a dead unit. In those cases the sub-second, session-preserving handover of CARP is worth the extra box and complexity.

I am not anti-HA; I am anti-mismatched-HA. There are homes where it makes real sense. If you run video calls all day for work, an update that reboots your only firewall mid-call is a genuine problem, and an HA pair lets you update one node at a time with zero downtime — which is honestly the feature I find most compelling, more than the hardware-failure case. If you self-host services family or friends rely on, seamless failover of the firewall keeps them up. And if your rack is somewhere you cannot reach in minutes, a live standby beats driving across town. The honest test: can you tolerate a few minutes of downtime to reboot or swap a firewall? If yes, skip HA. If genuinely no, it earns its keep.

Can You Have Both Failover and High Availability?

Yes, and for a truly critical home setup that is the complete answer: an HA firewall pair, each with multiple WANs, so you survive both a line outage and a box failure. This is belt and suspenders — the internet is redundant and the firewall is redundant. It is also the most complex and expensive home network you can build, and almost nobody needs it.

The layering is straightforward in concept: multi-WAN failover handles the internet line, CARP handles the firewall hardware, and together they cover every common failure mode. In practice, combining them multiplies the configuration surface — two firewalls to keep in sync, multiple gateways per box, careful CARP-plus-gateway-group interaction — and it is easy to introduce bugs that make you less reliable, not more. If you are going to do it, build and verify failover first on a single box, confirm it works with a real unplug test, and only then layer HA on top once the simpler piece is proven. The pfSense HA setup guide covers the clustering side. But be honest about whether you are solving a real problem or collecting complexity for its own sake — the second is a hobby, not a requirement.

So What Should You Actually Build?

Failover, almost certainly. Put your money and effort into a second internet line — a cellular backup is ideal — and a firewall that switches to it automatically. That single upgrade defends against the outage your home is genuinely likely to experience, and it does so at a fraction of the cost and complexity of high availability. Add a UPS so a power blip does not masquerade as a hardware failure, and keep a config backup for the rare day the box itself dies.

Reach for high availability only if you have named a specific, real need for sub-second firewall redundancy — a home office you cannot afford to drop, services others depend on, a rack you cannot reach. And if you need both, build failover first, prove it, then add HA on top. The instinct that got me here is the same one that runs through this whole site: match the tool to the actual failure, right-size for a home rather than pretending it is a datacenter, and never buy complexity you cannot justify. Start with the failover build and you will have solved the problem you actually have.

Keep Building

• Multi-WAN & Internet Failover: A Home Network Guide
• Setting Up Multi-WAN Failover in OPNsense
• Internet Failover With a 5G or LTE Backup Modem
• pfSense High Availability Setup
• UPS Integration for Proxmox and the Rack