AsusWRT-Merlin 380.59 AC68U Port Fwd Apply fail

[eyeks]

No matter what port configuration is entered to setup port forwarding on Asus RT-AC68U, running AsusWRT-Merlin v380.59, upon tapping Apply - all browsers report: Please enter a value between 1 and 65535. I tried Chrome 52.0.2743.82 m, Edge 25.10586.0.0, Opera Beta 39.0.2256.37 on 3 separate Windows 10 machines - all reported the same error. There was no way to apply the port forwarding configuration. See screen capture at: https://goo.gl/IjKQqu

I then proceeded to down-grade to 378.56_2. Upon reboot, I was able to apply the port forwards without any problems. Chime in if you've either seen or not seen this problem with 380.59.

 

[eyeks]

I then proceeded to down-grade to 378.56_2. Upon reboot, I was able to apply the port forwards without any problems. Chime in if you've either seen or not seen this problem with 380.59.

Came back an hour later and the port forward configurations were gone. When I try to re-apply them, I am once again getting the dreaded: "Please enter a value between 1 and 65535." Now on 378.56_2 - HELP! My remote backups and Adobe Acrobat collaboration are broken!!

 

[john9527]

Came back an hour later and the port forward configurations were gone. When I try to re-apply them, I am once again getting the dreaded: "Please enter a value between 1 and 65535." Now on 378.56_2 - HELP! My remote backups and Adobe Acrobat collaboration are broken!!

Check your nvram utilization on the tools page. If nvram is near full, port forwards seem to be the first thing affected.

 

[eyeks]

Check your nvram utilization on the tools page. If nvram is near full, port forwards seem to be the first thing affected.

Thank you, @john9527 - You hit the nail on the head! Why doesn't this show up on Google searches for this symptom? How do we seed the search for those facing it in future?

And, more importantly, how do I fix it?

 

[john9527]

And, more importantly, how do I fix it?

This is really just a matter of brute force investigation....

To start, you can run this command via telnet/ssh

nvram show | awk '{print length(), $0 | "sort -n -r"}' | cut -d"=" -f 1 | head -n 20

This will list the top 20 nvram vars by size. This may give you a hint as to what is consuming the nvram space. One thing to look for is vars containing the string 'clientlist', especially if you have a lot of clients coming and going on the network. If you do a search, you'll find some scripts that can periodically clean this out for you (when running Merlin firmware).

Otherwise, you'll need to do a factory default reset to get a baseline nvram used. Then start configuring things, checking the nvram utilization at each step. It's possible that you are just trying to use too many features simultaneously and overwhelming the router capability. At this point, it's decision time as to what you really need.

My rule of thumb is to keep nvram utilization at 90% or less.....by 95% utilization you are likely to start having all sorts of strange problems.

 

[eyeks]

Finally - a big thanks again @john9527 - you-the-man!

Unfortunately, I wasn't able to run the piped awk because the router deteriorated to the point it wouldn't allow a login. I ended up doing a factory reset and a from-scratch reconfig... keeping a close eye on the nvram usage - it's now hovering around 49k/64k.

With such a pervasive error with nvram becoming full - it would be good to have an alert when it's dangerously full and a nvram (top 10) usage report would have prevented my suffering!

Best wishes!
-eyeks

 

[RMerlin]

Having a webui notification would work, but it's not ideal as people don't necessarily log regularly onto their router.

Asus is working on a new notification center. I'm waiting to see where they are headed with that before deciding on implementing any monitoring method for current nvram usage.

 

[john9527]

Well, I went ahead and added an nvram utilitzation alert in my last fork release (I wasn't aware of the Asus work, but probably couldn't take advantage of it on my fork anyway).

Because of what Merlin said about people logging in the router, it may not be proactive. But in the event of a problem, people will eventually get to the router and then will be alerted.

 

[AppleBag]

If you do a search, you'll find some scripts that can periodically clean this out for you

Hey John

Do you have one of these scripts handy? I did some searching but didn't come up with anything (maybe I'm not 100% sure what to look for, I'm not very schooled in Linux).

I ran into this issue and came across this thread. After reading your post I went and took a look, and sure enough :

The blurred out stuff is actually some user feedback I sent a long time ago via the in-app feedback button, before I even switched from the stock firmware over to merlin, lol. Right now I'm running the current (as of right now) version 384.8_2 of merlin on a ASUS RT-AC68U.

I found the backup script that backs up/restores nvram/jffs settings, backed it all up, router reset, restored settings, and now I'm at a comfy 60661 / 65536 bytes, but I'd still like to have a script that does what you mention handy if and/or when I run into that problem again.

TIA

 

[john9527]

Let's check first what is using up your nvram space, and we can customize a script for you (there are side effects, such as possibly losing your networkmap customizations). I searched for the script I remembered somebody had put together, but wasn't able to find it.

log in to the router and run this command in post #5 of the thread to list your top 20 largest nvram settings.

 

[AppleBag]

Let's check first what is using up your nvram space, and we can customize a script for you (there are side effects, such as possibly losing your networkmap customizations). I searched for the script I remembered somebody had put together, but wasn't able to find it.

log in to the router and run this command in post #5 of the thread to list your top 20 largest nvram settings.

Hi John,

Thanks, the second screenshot in my previous post is the result of running that command.