WAX630 continuous reboot loop: dvlan_fallback=1 firmware bug (hostapd_tr SIGABRT)

Hardware: WAX630, firmware 12.5.0.15, managed via Netgear Insight

My WAX630 has been crashing and rebooting every 15–20 minutes for months, and after digging through the AP's own diagnostic logs I've pinpointed the root cause. Sharing here in case anyone else hits this and hoping someone has found a workaround I haven't tried yet.

What's happening

After every reboot, the crash repeats on the same ~17-minute cycle:

hostapd_tr crashes with SIGABRT (signal 6)

→ configd loses IPC connection

→ nddmp database timeouts

→ apmonitord watchdog fires ("confd_stuck_handler: No response from configd")

→ AP reboots

→ repeat

The AP's own `/etc/sysconfiglog/log/` directory has three `core.hostapd_tr.*.tar` crash dumps spanning March 2025 through July 2025, all with the same signal-6 signature, so this has been happening across multiple firmware versions (V10.8.12.9 and V10.8.13.2, and now current 12.5.0.15).

Root cause (from the logs)

The firmware config generator is writing `dvlan_fallback=1` on every VAP even though Dynamic VLAN (`dvlan_status`) is `0` (disabled) on every SSID. That contradiction means hostapd tries to open an L2 packet interface on the VLAN bridge (`br-lan`) for each VAP on every boot, and fails. Every boot produces 193 occurrences of this error in `hostapd.txt`:

Failed to open l2_packet interface for vlan bridge

Those failures flood hostapd_tr until it hits an unrecoverable state and aborts.

The fix is straightforward: set `dvlan_fallback=0` on all VAPs. But `dvlan_fallback` isn't exposed anywhere in the Insight UI, and SSH is locked to Netgear engineering on Insight-managed APs, so there's no way for an end user to push that change directly.

Proof it's WAX630-specific, not a config issue

I compared against a WAX610Y at the same Insight location receiving the identical config push:

Same Insight account, same location, same SSIDs, same config, zero errors on the WAX610Y, constant crashes on the WAX630. This is a WAX630-specific hostapd behavior, not a network or config mistake on my end from what I can tell.

What I've already tried (all failed to fix dvlan_fallback)

• Toggling VLAN IDs (1 → 2 → 1) on all SSIDs via Insight
• Enabling WPA2 Enterprise + Dynamic VLAN on a test SSID, then disabling `dvlan_fallback` still written as `1`
• SSH resets at KEXINIT, confirmed blocked for end users on Insight-managed APs
• Factory reset + re-adoption, reset firmware on the device to pick up net-new configuration from Insight, but since `dvlan_fallback=1` is written by the config generator on every Insight push, it reproduces the same config

Temporary stability workaround (not a fix)

I stabilized the AP by removing the secondary crash trigger: a client device was connecting at -88 dBm RSSI in a rapid 802.11r roaming loop, creating additional hostapd churn on top of the dvlan_fallback stress. Changes applied via Insight:

• Disabled 802.11r (Fast Roaming) location-wide
• Enabled Load Balancing → Balance on Client Rx RSSI (value 23 ≈ -72 dBm threshold)
• Enabled Disassociate Sticky Clients (2.4GHz: -70 dBm, 5GHz: -80 dBm, 6GHz: -78 dBm)

After these changes: 12+ hours continuous uptime, zero l2_packet errors on current boot. But `dvlan_fallback=1` is still in the config; this is a bandage, not a fix. If hostapd stress increases for any reason, the underlying bug will reassert.

Netgear's own 12.5.0.15 release notes list "Hostapd interface wifixvapx is unresponsive" as a known open issue.  This appears to be the specific config path (dvlan_fallback=1 + dvlan_status=0) that triggers it.

What I'm looking for

1. Has anyone found a way to force `dvlan_fallback=0` through the Insight UI (or any other end-user-accessible method)?
2. Has anyone had Netgear support successfully push a remote config fix for this?
3. Is anyone else seeing this same crash pattern on WAX630? Would be helpful to know if this is widespread.

I have detailed logs including three device-generated core dumps and before/after comparisons if anyone from Netgear engineering wants to look. Happy to share.

Firmware: 12.5.0.15 | Management: Insight | Crash signal: SIGABRT (6) | Affected process: hostapd_tr