UniFi OS 5.1.11 on the UCG Fiber has a critical bug in firewall rule configuration. Enabling logging on a DEFAULT DROP rule silently breaks the entire inter-zone filter chain for that zone. The UI shows all rules intact. The kernel shows an empty chain. And the only way to discover this is to SSH into the gateway and read the actual iptables state — by then, traffic that should be blocked has been flowing for hours.
The Bug: DEFAULT DROP Rule Logging Corrupts the Configuration#
On UniFi OS 5.1.11 with Network 10.4.57, enabling the logging toggle on any DEFAULT DROP rule triggers a configuration migration event. UniFi validates the new rule object and discovers a schema mismatch: the FirewallRuleLog field expects an object with specific properties, but the schema validation is receiving a boolean instead. The validation fails silently. The rule is not marked invalid in the UI. The system does not alert the admin. Instead, UniFi skips pushing the rule to the kernel's iptables chain, severing the entire udapi-to-iptables pipeline for that zone's filtering.
The result: a zone chain that appears to have 15 rules in the UniFi UI contains zero rules in the kernel.
Symptom: Traffic Passes When It Should Be Blocked#
The first symptom is traffic arriving on a subnet where it should be denied. A device on the Skynet VLAN reaches a service on the Admin VLAN. A remote user on the Users VLAN connects to something on MGMT. The firewall logs show no block. The rules appear correct in the UI. But the traffic is not being filtered — it's flowing straight through.
The second symptom is checking UniFi's event logs. You see firewall rule updates timestamped around the time you enabled logging. You don't see any error. You assume the update succeeded. It did not.
Discovery: SSH and Read the Kernel State#
The UI is not authoritative. The kernel is. You must verify every firewall change by SSHing into the gateway and reading iptables directly.
SSH to the UCG Fiber via the admin interface. You'll be prompted for the UNIFI_OS_PASSWORD (not your UniFi account password — the OS root password, separate from the web UI).
Once connected, list the rules in a specific zone chain. UniFi chains follow the pattern UBIOS_{SRC}_{DST}_USER. Internal traffic to WAN is UBIOS_LAN_WAN_USER. Trusted traffic to WAN is UBIOS_CUSTOM1_WAN_USER. Untrusted traffic to LAN is UBIOS_CUSTOM2_LAN_USER. VPN traffic to Internal is UBIOS_VPN_LAN_USER.
Run: iptables -L UBIOS_LAN_CUSTOM1_USER -n --line-numbers. The -n flag skips DNS resolution (faster), --line-numbers shows each rule's position for cross-referencing with the UI.
If the chain has zero rules, you've found the bug.
Chain Naming and Zone Type Codes#
UniFi generates chain names using zone type codes: Internal = LAN, Untrusted = CUSTOM2, Trusted = CUSTOM1, VPN = VPN. Learn these, bookmark them. You'll be reading a lot of kernel rules.
The ipset Paradox: Broken Groups Can't Be Recovered#
When you delete and recreate a rule using the same IP group (the same set of CIDRs), UniFi reuses the underlying kernel ipset. It hashes the group by its content. If the ipset was corrupted by the original rule creation failure, the new rule will attach to the same broken ipset.
A list:set parent ipset can return NOT IN SET even though the child hash:net contains the exact CIDR you're checking. This indicates broken kernel state. Recreating the rule in the UI will not fix it — UniFi will hash the content identically and reattach to the same broken set.
The only recovery is to create a new IP group with a different name. Add a version number or timestamp suffix. This forces UniFi to generate a new ipset in the kernel. Then recreate your rules using the new group.
The Non-Fixes#
Restarting unifi.service does not fix this bug. Rebooting the gateway does not fix it. Neither does re-running the UniFi controller's migration tasks or updating Network versions.
The only fix is to delete every affected rule one by one and recreate them from scratch. If the underlying ipsets are broken, create new IP groups so UniFi generates fresh ipsets. It's tedious. It's necessary.
Prevention: Never Enable Logging on DEFAULT DROP#
Do not enable the logging toggle on DEFAULT DROP rules. If you need to verify that drops are occurring, use iptables -L -v and read the hit counters — the [packet count] column shows how many packets have matched each rule. Or check the UniFi traffic logs from the UI without enabling per-rule logging on the DEFAULT rule itself.
Bug Status and Verification Workflow#
This bug was submitted to Ubiquiti on May 19, 2026. As of June 4, 2026, it remains unresolved on UniFi OS 5.1.11.
After every firewall change, run the verification workflow. SSH into the gateway. List each affected zone chain. Count the rules — they must match what the UI shows. Then use ipset list to recursively check that your IP groups reference valid child ipsets three levels deep. Silence from the UI is not confirmation. The kernel is the truth.