WiFi troubleshooting guide, fix empty security overrides, update config docs

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
This commit is contained in:
jazzymc
2026-03-12 16:18:40 +02:00
parent 6320c0f8d9
commit d0b4fae25e
4 changed files with 265 additions and 8 deletions

View File

@@ -68,6 +68,7 @@ infrastructure/
├── 08-DNS-ARCHITECTURE.md # DNS failover architecture
├── 09-TAILSCALE-VPN.md # Tailscale VPN setup
├── 10-FAILOVER-NOBARA.md # VRRP failover to Nobara
├── 12-WIFI-TROUBLESHOOTING.md # WiFi/CAPsMAN troubleshooting guide
├── CHANGELOG.md # Change history
├── archive/ # Completed/legacy docs
│ └── vlan-migration/ # VLAN migration project artifacts

View File

@@ -1,6 +1,6 @@
# WiFi and CAPsMAN Configuration
**Last Updated:** 2026-02-26
**Last Updated:** 2026-03-12
**Purpose:** Document WiFi network settings, CAPsMAN configuration, and device compatibility requirements
---
@@ -23,11 +23,12 @@
| SSID | XTRM |
| Band | 5GHz |
| Mode | 802.11ax (WiFi 6) |
| Channel | 5180 MHz (ch 36) |
| Width | 40MHz |
| Channel | 5745 MHz (ch 149) |
| Width | 20/40/80MHz |
| Security | WPA2-PSK + WPA3-PSK |
| Cipher | CCMP (AES) |
| 802.11r (FT) | Enabled |
| 802.11r (FT) | Disabled |
| Skip DFS | All |
| Password | `M0stW4nt3d@home` |
---
@@ -44,12 +45,14 @@ Some devices (Tuya JMWZG1 gateway, Amazfit TREX3, iPad 2) require legacy setting
|---------|-------|--------|
| SSID | XTRM2 | |
| Band | 2.4GHz | IoT compatibility |
| Mode | **802.11g** | Legacy device support |
| Mode | **802.11g** | Legacy device support (NOT 802.11n — breaks IoT) |
| Channel | **1 (2412 MHz)** | Most compatible |
| Width | **20MHz** | Required for old devices |
| Security | **WPA-PSK + WPA2-PSK** | WPA needed for legacy |
| Cipher | **TKIP + CCMP** | TKIP required for old devices |
| 802.11r (FT) | **Disabled** | Causes issues with IoT |
**CRITICAL:** Security must be set explicitly on the interface, not just the profile. Empty `security.authentication-types=""` means OPEN network, not "inherit from profile." See [12-WIFI-TROUBLESHOOTING.md](12-WIFI-TROUBLESHOOTING.md).
| Password | `M0stW4nt3d@IoT` | |
### Devices Requiring WPA + TKIP
@@ -140,10 +143,10 @@ The CAP runs bridge VLAN filtering to properly tag/untag WiFi client traffic bef
| Interface | Radio | Band | SSID | Security | Status |
|-----------|-------|------|------|----------|--------|
| cap-wifi1 | wifi2 | 5GHz | XTRM | WPA2/WPA3-PSK, CCMP | Working (Ch 52/5260, 40MHz, DFS) |
| cap-wifi2 | wifi1 | 2.4GHz | XTRM2 | WPA2-PSK, CCMP | Working (Ch 6/2437, 20MHz) |
| cap-wifi1 | MAC :BE | 2.4GHz | XTRM2 | WPA+WPA2, TKIP+CCMP | Working (Ch 13/2472, 20MHz, 802.11g) |
| cap-wifi2 | MAC :BF | 5GHz | XTRM | WPA2/WPA3-PSK, CCMP | Working (Ch 36/5180, 20/40/80MHz, 802.11ac) |
**Note:** cap-wifi2 uses WPA2+CCMP only (not WPA+TKIP like HAP's local wifi2). Legacy IoT devices requiring TKIP will only work on HAP1's local wifi2.
**Note:** CAP radios swapped after CAPsMAN re-provisioning. Identify by MAC address, not interface name. See [12-WIFI-TROUBLESHOOTING.md](12-WIFI-TROUBLESHOOTING.md) for re-provisioning procedures.
---

View File

@@ -0,0 +1,237 @@
# WiFi / CAPsMAN Troubleshooting Guide
**Last Updated:** 2026-03-12
**Purpose:** Document known pitfalls, root causes, and diagnostic procedures for WiFi and CAPsMAN issues on the MikroTik HAP ax³ + cAP XL ac setup.
---
## Pitfall 1: Empty Inline Security Overrides = Open Network
**Severity:** CRITICAL
**Problem:** Setting `security.authentication-types=""` on a WiFi interface does NOT mean "inherit from security profile." RouterOS interprets an empty string as **no authentication (open network)**.
**Symptoms:**
- Devices connect but show empty AUTH-TYPE in registration table
- IoT devices that try WPA/WPA2 handshake silently fail — router logs show ZERO connection attempts
- Other devices connect fine (they accept open)
**Root Cause:** Attempting to "clear inline overrides" to inherit from the security profile by setting empty values. RouterOS treats empty string as explicit "no auth."
**Fix:**
```routeros
# Always set explicit values on the interface
/interface wifi set wifi2 \
security.authentication-types=wpa-psk,wpa2-psk \
security.encryption=tkip,ccmp
```
**Rule:** NEVER set `security.authentication-types=""` or `security.encryption=""`. Always use explicit values matching the security profile.
---
## Pitfall 2: CAPsMAN Re-Provisioning Wipes Interface Config
**Severity:** HIGH
**Problem:** Running `/interface wifi capsman remote-cap provision` clears all configuration from cap-wifi interfaces — security, channel, datapath, and SSID are all removed. Interfaces show "SSID not set" and remain inactive.
**Fix:** After re-provisioning, manually re-apply full config:
```routeros
# 2.4GHz (cap-wifi1 = MAC :BE = 2.4GHz radio)
/interface wifi set cap-wifi1 \
configuration=cfg-xtrm2 security=sec-xtrm2 datapath=dp-cap \
channel.frequency=2472 channel.band=2ghz-g channel.width=20mhz
# 5GHz (cap-wifi2 = MAC :BF = 5GHz radio)
/interface wifi set cap-wifi2 \
configuration=cfg-xtrm security=sec-xtrm datapath=dp-cap \
channel.frequency=5180 channel.band=5ghz-ac \
channel.width=20/40/80mhz channel.skip-dfs-channels=all
# Re-enable both
/interface wifi enable cap-wifi1
/interface wifi enable cap-wifi2
```
---
## Pitfall 3: Interface IDs Change After Re-Provisioning
**Severity:** HIGH
**Problem:** After CAPsMAN re-provisioning, cap-wifi interface internal IDs change (e.g., `*20`/`*21` become `*22`/`*23`). Access-list rules referencing old IDs stop matching.
**Symptom:** `client was disconnected because could not assign vlan` error on CAP interfaces.
**Fix:**
```routeros
# Check current IDs
:foreach i in=[/interface wifi find where name~"cap"] do={
:put ([/interface wifi get $i name] . " = " . $i)
}
# Update access-list rules
/interface wifi access-list set [find where interface=*OLD] interface=*NEW
```
**Best practice:** Don't use CAP-specific access-list rules. Let all clients (HAP and CAP) use the same MAC-based access list. The HAP handles VLAN assignment uniformly via CAPsMAN.
---
## Pitfall 4: CAP Radio-to-Interface Mapping Swap
**Severity:** MEDIUM
**Problem:** After re-provisioning, `cap-wifi1` and `cap-wifi2` may swap which physical radio they map to. Assigning 5GHz config to the 2.4GHz radio (or vice versa) causes "no available channels" error.
**Identification:** Check MAC addresses:
| MAC suffix | Radio | Must receive |
|------------|-------|--------------|
| :BE | 2.4GHz | 2.4GHz config (XTRM2) |
| :BF | 5GHz | 5GHz config (XTRM) |
**Fix:** Match config to the correct radio MAC, not the interface name.
---
## Pitfall 5: CAP Band Must Be AC, Not AX
**Severity:** MEDIUM
**Problem:** The cAP XL ac only supports 802.11ac. Setting band to `5ghz-ax` results in the radio not starting.
**Fix:** Always use `5ghz-ac` for the CAP 5GHz channel configuration.
---
## Pitfall 6: IoT Devices Need Legacy WiFi Settings
**Severity:** HIGH
**Problem:** Many IoT devices (vacuums, smart gateways, ovens) require legacy WiFi settings to connect. Using 802.11n-only or WPA2-only silently prevents connections — the router sees zero attempts.
**Required XTRM2 (2.4GHz) settings:**
| Setting | Value | Reason |
|---------|-------|--------|
| Band | `2ghz-g` | NOT `2ghz-n` — IoT devices may only support 802.11g |
| Auth | `wpa-psk,wpa2-psk` | Some devices need WPA1 available |
| Encryption | `tkip,ccmp` | Some devices need TKIP |
| Channel width | `20mhz` | Maximum compatibility |
| FT (802.11r) | Disabled | Causes issues with IoT |
**Known devices requiring legacy support:**
- Roborock S7 Vacuum (B0:4A:39:3F:9A:14)
- Tuya Smart Gateway JMWZG1 (38:1F:8D:04:6F:E4)
- Bosch Oven (94:27:70:1E:0C:EE)
- Various other IoT appliances
---
## 5GHz Channel Separation
HAP and CAP must use different 5GHz channels to avoid co-channel interference:
| Device | Channel | Frequency | Band |
|--------|---------|-----------|------|
| HAP wifi1 | 149 | 5745 MHz | 5ghz-ax |
| CAP cap-wifi2 | 36 | 5180 MHz | 5ghz-ac |
Both use `skip-dfs-channels=all` to avoid radar detection disconnects.
---
## Diagnostic Checklist
When devices can't connect to WiFi, check in this order:
### Step 1: Check Security (Most Common Issue)
```routeros
# Check if AUTH-TYPE is empty in registration table (= open network!)
/interface wifi registration-table print
# Check inline security overrides
:put [/interface wifi get wifi2 security.authentication-types]
:put [/interface wifi get wifi2 security.encryption]
# If empty → security is broken, set explicit values
```
### Step 2: Check Band Compatibility
```routeros
/interface wifi monitor wifi2 once
# If channel shows /n → change to /g for IoT compatibility
```
### Step 3: Enable Debug Logging
```routeros
/system logging add topics=wireless,debug action=memory
# Then check: /log print where topics~"wireless"
```
### Step 4: Check CAP Interface IDs
```routeros
# Verify access-list rules reference current IDs
:foreach i in=[/interface wifi find where name~"cap"] do={
:put ([/interface wifi get $i name] . " = " . $i)
}
/interface wifi access-list print where interface~"\\*"
```
### Step 5: Check Radio-MAC Mapping
```routeros
# Verify cap interfaces are assigned to correct radios
/interface wifi print where name~"cap" proplist=name,mac-address,channel.band
```
### Step 6: If Router Sees ZERO Attempts
This means:
- **Security mismatch** — device won't even try (most likely empty auth = open)
- **Band incompatibility** — 802.11n-only blocks 802.11g devices
- **Device-side issue** — power cycle device, re-do WiFi setup from scratch
---
## Quick Recovery Commands
### Restore XTRM2 (2.4GHz) to known working state
```routeros
/interface wifi security set sec-xtrm2 \
authentication-types=wpa-psk,wpa2-psk encryption=tkip,ccmp
/interface wifi set wifi2 \
security.authentication-types=wpa-psk,wpa2-psk \
security.encryption=tkip,ccmp \
security.ft=no security.ft-over-ds=no
/interface wifi channel set ch-2g-hap \
frequency=2412 band=2ghz-g width=20mhz
```
### Restore XTRM (5GHz) to known working state
```routeros
/interface wifi security set sec-xtrm \
authentication-types=wpa2-psk,wpa3-psk encryption=ccmp
/interface wifi set wifi1 \
security.authentication-types=wpa2-psk,wpa3-psk \
security.encryption=ccmp \
security.ft=no security.ft-over-ds=no
/interface wifi channel set ch-5g-hap \
frequency=5745 band=5ghz-ax width=20/40/80mhz skip-dfs-channels=all
```
### Restore CAP interfaces after re-provisioning
```routeros
/interface wifi set cap-wifi1 \
configuration=cfg-xtrm2 security=sec-xtrm2 datapath=dp-cap \
channel.frequency=2472 channel.band=2ghz-g channel.width=20mhz
/interface wifi set cap-wifi2 \
configuration=cfg-xtrm security=sec-xtrm datapath=dp-cap \
channel.frequency=5180 channel.band=5ghz-ac \
channel.width=20/40/80mhz channel.skip-dfs-channels=all
/interface wifi enable cap-wifi1
/interface wifi enable cap-wifi2
```

View File

@@ -2,6 +2,22 @@
**Purpose:** Major infrastructure events only. Minor changes are in git commit messages.
---
## 2026-03-12
### WiFi Optimization & Troubleshooting
- **[WIFI]** Moved HAP 5GHz from ch 36 (5180) to ch 149 (5745), skip-dfs-channels=all
- **[WIFI]** Moved CAP 5GHz from ch 52 (5260) to ch 36 (5180), band corrected from ax to ac
- **[WIFI]** Separated HAP/CAP 5GHz channels to avoid co-channel interference
- **[WIFI]** Fixed sec-xtrm2 security: WPA+WPA2 with TKIP+CCMP for IoT compatibility
- **[WIFI]** Fixed critical bug: empty inline security.authentication-types="" caused wifi2 to run as open network — IoT devices silently failed to connect
- **[WIFI]** Set explicit encryption on all interfaces and security profiles (never leave empty)
- **[WIFI]** Removed CAP-specific access-list catch-all rules — all clients now use unified MAC-based access list
- **[WIFI]** Fixed CAP interface IDs in access-list after re-provisioning (*20/*21 → *22/*23)
- **[WIFI]** Restored 2.4GHz band to 2ghz-g (was changed to 2ghz-n, breaking IoT devices)
- **[WIFI]** Disabled FT (802.11r) on wifi1 (5GHz) for stability
- **[DOCS]** Added 12-WIFI-TROUBLESHOOTING.md with diagnostic checklist and recovery commands
---
## 2026-02-28