AC68U UI Crashing occasionally

[ccjj]

Hi all,

Im a massive fan of my AC68U and thanks to Merlin for the epic firmware (running 384.19 latest version). However Im having some problems with it and finally decided its time to investigate a bit more....

Every now and again (approx once a month) the web interface just stops responding (does not load a page), and the SSH connection is the same (sometimes it will show a login prompt but then when you attempt to login it spins continuously and then times out). Despite this, the router continues to operate totally normally (WiFi/Diversion adblock/VPN tunnel/port forwards), I just cant do anything new or check anything. (I'll call this the zombie state)

A restart of the router always fixes things until it happens next time.... One of the most common situations I have noticed this is when logging into the router I type my credentials and something doesn't work, then from this point onward it reaches zombie state. I have added weekly reboots, but this only helps a little since sometimes it can happen quickly (e.g. just did it after 2 days, not sure if the reboot will now happen).

Either way Im a bit stuck, when I reboot there is nothing in the event log that looks suspicious and Im not sure how to debug. I have reset the NVRAM on 384.18 and then loaded the saved config file (rather than re-entering from scratch), but have not done it on 384.19. I dont think Im running anything super complex, the only addition from standard Merlin firmware is Diversion.

Please can I have some advice/recommendations?

Thanks in advance!

 

[ColinTaylor]

I've seen numerous posts/threads reporting this same problem. Maybe there is some advice in one of those.

 

[ccjj]

Thanks, yes I've found several of those, but mostly they seem to say you can just reboot httpd and it fixes things, however I can't even log into SSH when it goes into zombie mode. I'm thinking this might be something different, but don't know for sure, and was hoping there was some way of living what was happening so I could debug a bit more...
Thanks

 

[sHORTYWZ]

Just actually came here to report the exact same thing, as I sit and watch the spinning circle trying to load my page.

The login screen is responsive, but once I enter my credentials, nothing will load.

 

[ccjj]

Just actually came here to report the exact same thing, as I sit and watch the spinning circle trying to load my page.

The login screen is responsive, but once I enter my credentials, nothing will load.

And just to confirm once this has happened there is no response from the router until next power cycle? On web or SSH?

 

[sHORTYWZ]

And just to confirm once this has happened there is no response from the router until next power cycle? On web or SSH?

Correct... ssh poses a similar sequence of events:

admin@192.168.1.1's password:

ASUSWRT-Merlin RT-AC68U 384.19_0 Fri Aug 14 19:17:44 UTC 2020

Cannot get past that point - never receive a prompt.

 

[sHORTYWZ]

Just want to circle back here - I stumbled upon a post somewhere else (which I unfortunately cannot find anymore) that pointed to this being a potential issue with the auto channel selection for the WAN.

I disabled this on both WAN networks, and I have not experienced the above issue since - 3 weeks of uptime and everything is still responsive - would have generally required at least a couple restarts by now.

 

[ccjj]

That's interesting, I've been giving it a go with the auto logout timer set at 1 hour, rather than unlimited and has been stable so far with that (I think 2 weeks). For good measure I'll give that a go too, and if it's definitely fixed I can happily do the tests individually to see which one worked.

EDIT does that mean you removed auto from the control channel or control and extension? I'm not sure how it will affect performance of I change both, but hopefully not too much...

Thanks

 

[sHORTYWZ]

EDIT does that mean you removed auto from the control channel or control and extension? I'm not sure how it will affect performance of I change both, but hopefully not too much...

Extension was left to Auto, as there is no way (at least not for me) to move away from that.

 

[ccjj]

@sHORTYWZ - I have made this change and was running well for almost a month, but sadly the webui has crashed again now, so needing a router restart. Did you manage more than a month?

 

[Blee]

same issue in RT-AC86U... have to reboot and it's ok

 

[sHORTYWZ]

@ccjj still running strong since I made the change - not a single reboot since.

 

[HarryH3]

Another thread mentioned that turning off auto-checking for updates solved the problem. You'll find it at Administration -> Firmware Upgrade tab -> Scheduled check for new firmware availability. Set it to No. Mine would lose its mind pretty often but it's been great for the last week to week and a half since I changed that setting.

 

[HarryH3]

Another thread mentioned that turning off auto-checking for updates solved the problem. You'll find it at Administration -> Firmware Upgrade tab -> Scheduled check for new firmware availability. Set it to No. Mine would lose its mind pretty often but it's been great for the last week to week and a half since I changed that setting.

And this morning, no response once again. Had to power cycle to get to the home page.

 

[ccjj]

Im glad its not just me, happened twice in a week this week. I cant for the life of me work out what the trigger is....

What custom services are you using on your router?

The only "complex" things Im running are:
* DDNS updater (alongside Letsencrypt, though never got that working with my custom domain)
* VPN server (not actually used as I cant get the SSL cert working)
* VPN client (not connected though so not really used)
* Diversion ad-blocker (and associated pixelserv)
* Traffic monitor stats saved to USB disk (nothing else using USB)

Guest Wifi disabled, game services disabled, AI cloud stuff disabled, IPv6 fully disabled

EDIT: This time around I could access SSH when it crashed (again I have no idea what was different, but after a couple of attempts at SSH it let me in). However most commands failed, e.g. typing 'reboot' gave me "mmap of a spare page failed!" as an error message, and typing 'diversion' gives "sh: diversion: not found"
EDIT2: After restarting the router I have discovered that Diversion is no longer installed, I guess that probably happens if you wipe JFFS or something? Either way I have now reinstalled it

 

[HarryH3]

Mine just runs a VPN server and saves traffic stats to a USB flash drive, and has guest wifi enabled.

 

[ccjj]

I have a strange feeling it is somehow related to the traffic stats, as I noticed the traffic stats are blank when the router stops responding, perhaps some kind of cache is corrupted and it fails to run a certain service....? We need somebody else to comment and see if they have traffic stats enabled....!