Randomly drop into recovery after flash merlin

[Sylphia]

Hi, I've bought a AC66U few days ago, and upgraded to lastest stock firmware. Then I flashed merlin(374.43) through web ui and successfully log in to merlin version, but when I soft reboot by web ui, it just drop me in to recovery page, and the clear nvram / reboot button does no effect. After n times' hard reboot, I got a working recovery page and cleared nvarm, rebooted, and it booted into web ui, from there I reflashed stock and reboot with no issue. Then I tried again to flash merlin, it seemed all ok at first, rebooted many times with no issue, then after another rebooting, I found wifi broken, 2.4GHz was on channel 149, 5GHz pops a dialog said Country code invalid, no channel selection available, so I rebooted again and yeah, recovery mode... cleared nvarm, reboot, reflash stock...

Do I need to RMA?

 

[RMerlin]

For some reason, I've seen a few identical issues being reported over the past two years, every time with the RT-AC66U. I have to wonder if maybe this wouldn't be a low-level issue with the RT-AC66U that only happens to a small group of users.

In any case, nobody ever found any reason why this happened to a handful of people, and always only with the RT-AC66U. The only way to troubleshoot this would be to be able to boot with a serial cable hooked to the PCboard, and see if there's anything during the reboot indicating as to why the router believes it needs to boot in recovery mode.

 

[Sylphia]

Thank you for quick response and great work.

It appears rebooting cause some data loss (nvram?), i.e. the wifi country code I've mentioned, and became more worse at next reboot, and then lead to boot failure.

Or, I'm not very sure it's a boot time failure, more likely something went bad when saving configuration.
As I remember, before reboot to death, the last "apply" action leads to a page of "Your router config is changed" or something similar to that. Then 2 situations happen: 1.refresh and the "Your router config is changed" forever, reboot to death. 2. refresh to normal web ui, but something weird shows up, in my case, wifi config messed up, then reboot to death.

For now, stock firmware seems working fine, hope the cause can be found and fixed.

 

[Sylphia]

I think I've reproduced this particular issue.

1. Flash merlin firmware (374.43).
2. Do initital settings.
3. Create a OpenVPN client. In my case, I imported a .ovpn config file, and set server address, port, Extra HMAC authorization, user name, password, then switch the client On.
4. OpenVPN client will not running cause "Jun 27 23:01:22 openvpn[568]: Options error: No client-side authentication method is specified. You must use either --cert/--key, --pkcs12, or --auth-user-pass", although I've entered username/password before switch client on - could be a bug, it's not important for this case.
5. Enter username/password again for OpenVPN client and switch it on, this time it should succeed.
6. Edit 2.4GHz wifi setting, click apply, done. Now my config messed up, 2.4GHz running on channel 149, 5 GHz disabled cause no valid channel: "No country code".

Below is part of my log:
___________________________________________________

3:28 openvpn[582]: WARNING: No server certificate verification method has been enabled. See http://openvpn.net/howto.html#mitm for more info.
Jun 27 23:03:28 openvpn[582]: Attempting to establish TCP connection with [AF_INET]124.248.205.47:443 [nonblock]
Jun 27 23:03:29 openvpn[582]: TCP connection established with [AF_INET]124.248.205.47:443
Jun 27 23:03:29 openvpn[582]: TCPv4_CLIENT link local: [undef]
Jun 27 23:03:29 openvpn[582]: TCPv4_CLIENT link remote: [AF_INET]124.248.205.47:443
Jun 27 23:04:24 rc_service: httpd 369:notify_rc restart_wireless
Jun 27 23:04:30 openvpn[582]: TLS Error: TLS key negotiation failed to occur within 60 seconds (check your network connectivity)
Jun 27 23:04:30 openvpn[582]: TLS Error: TLS handshake failed
Jun 27 23:04:30 openvpn[582]: Fatal TLS error (check_tls_errors_co), restarting
Jun 27 23:04:30 openvpn[582]: SIGUSR1[soft,tls-error] received, process restarting
Jun 27 23:04:30 kernel: device eth1 left promiscuous mode
Jun 27 23:04:30 kernel: br0: port 2(eth1) entering disabled state
Jun 27 23:04:30 kernel: device eth2 left promiscuous mode
Jun 27 23:04:30 kernel: br0: port 3(eth2) entering disabled state
Jun 27 23:04:33 kernel: wl_module_init: passivemode set to 0x0
Jun 27 23:04:33 kernel: wl driver 6.30.163.2002 (r382208) failed with code 21
Jun 27 23:04:33 kernel: eth1: Broadcom BCM4360 802.11 Wireless Controller 6.30.163.2002 (r382208)

__________________________________________

I've noticed a wireless driver fail in the log, but I think it was caused by bad configuration values, and after wifi broken, in Tools->Other settings, UDP Timeout: Assured's value became undefined, no value for UDP Timeout: Unreplied, it's empty.

If I reboot the router, it will boot into recovery or simply only "Settings changed" page, a factory reset before rebooting is required to avoid this happening.

From what I've seen, I think this should be a software bug, not something low-level.
Could any one try this out?