RT-AC68P: cannot modify settings (EDIT: browser issue. See my post of 2015-12-18)

[David Arnstein]

I use the web interface to RT-AC68P. Today, I cannot modify any router settings. Specifically, when I press the "apply" button on any web page, the router presents the login screen. In particular, I cannot enable ssh access. This makes troubleshooting more difficult.

I experience this behavior with both 3.0.0.4.378.9135 and 3.0.0.4.378.9313. The only clue I have is that once and once only, I found this message in the general log:
Nov 14 11:19:07 dnsmasq-dhcp[430]: failed to write /var/lib/misc/dnsmasq.leases: No space left on device (retry in 60s)

I was only able to upgrade to .9313 by

1. Physical reset button on router
2. Minimal configuration through web interface
3. Upgrade firmware through web interface
4. Restore settings from a saved .CFG file, through web interface.

This suggests that there is something screwy about my saved settings. I am thinking about re-doing the entire configuration from scratch, but I hesitate. After I get all my settings typed in, I am concerned that I will encounter exactly the same error condition.

Any suggestions?

 

[RMerlin]

Your router is running out of memory. This issue should normally be resolved with 9313.

 

[David Arnstein]

This issue should normally be resolved with 9313.

Believe me, it is not.

 

[L&LD]

Believe me, it is not.

Possibly because you are using a backup cfg file to upload your settings with?

 

[David Arnstein]

Your router is running out of memory. This issue should normally be resolved with 9313.

Which memory? RAM or flash? The reason I ask is that I would upgrade to RT-AC3100 if it would help. The RT-AC68P and RT-AC3100 have the same amount of flash: 128 MBytes. But the RT-AC3100 has more RAM: 512 MBytes versus 256 MBytes.

 

[RMerlin]

Which memory? RAM or flash? The reason I ask is that I would upgrade to RT-AC3100 if it would help. The RT-AC68P and RT-AC3100 have the same amount of flash: 128 MBytes. But the RT-AC3100 has more RAM: 512 MBytes versus 256 MBytes.

RAM. It's caused by a bug in recent firmware releases by Asus where they left some debug logging active in RAM, causing routers to gradually exhaust their RAM.

 

[msb]

RAM. It's caused by a bug in recent firmware releases by Asus where they left some debug logging active in RAM, causing routers to gradually exhaust their RAM.

I'm assuming that a fix is in the works? Is it fixed in yours?

 

[RMerlin]

I'm assuming that a fix is in the works? Is it fixed in yours?

Fixed on Github.

 

[sm00thpapa]

I'm running .9313 on my 68P with no issues. I do reset to factory specs after each firmware update.

 

[David Arnstein]

Folks,

I spent most of Sunday on this issue. I repeatedly did the 10-second factory reset on the router, followed by a long session of modifying router settings via the web interface. I modified items such as router name and address, router password, wireless SSID and passcode, port forwarding rules, remote logging destination, and more.

Each time I tried, I would eventually arrive at the state where further modifications to router settings would become impossible. I was not able to find a particular router setting that triggered the failure. I mean, after each 10-second reset, I entered router settings in a different order, so that after each reset, the last router setting change before failure was different.

Is there something else I should do that is more cleansing than the 10-second reset? I really think that the router is physically defective. A bad patch in DRAM or flash, perhaps.

Thanks for any suggestions.

 

[Puppa]

Don't plug in your hard disk until after you've done most of your settings. You can reboot after applying just a few settings to avoid exhausting RAM. I take it you are on .9313 already?

Sent from my SM-N920V using Tapatalk

 

[David Arnstein]

Don't plug in your hard disk until after you've done most of your settings. You can reboot after applying just a few settings to avoid exhausting RAM. I take it you are on .9313 already?

This problem existed in -9313 and before. On Sunday, I did my testing with -1031.

I don't plug in any USB attachments into my router. Is that what you mean by "hard disk?"

I doubt that this is a RAM exhaustion problem. The problem begins to appear after a rather small set of configuration changes. I have not seen similar problems reported in these forums. If you have, please post!

 

[msb]

I really think that the router is physically defective.

Thanks for any suggestions.

I'd return it, I've never experienced anything remotely similar with mine...

 

[Puppa]

@David Arnstein Yes, by "hard disk" I was referring to USB, which has been a RAM-eater in the past. Does the router fall over even if you just let it sit there, powered on? (I suspect that it will.)

It's hard to say what exactly is wrong with it, but I think you're right: it's defective. I dusted off my RT-AC68U last night that was running a very old Merlin firmware and did a series of firmware upgrades to finally get to .1031, where I played around with more than a few settings. I didn't see even a hint of it not accepting changes. I did need to do a mini CFE boot (recovery mode) and press the clear NVRAM button at one point (did you try this, too?), but after that everything was golden.

 

[David Arnstein]

This is a browser issue. All three browsers on my main computer reproduce the issue. But when I switched to an Ubuntu Linux computer that I use rarely, the issue did not occur.

I will attempt to isolate the browser issue (issues?). If I learn anything, I will post a new thread.

Thanks for your help folks.