[Release 382] Asuswrt-Merlin 382.1 is now available

[RMerlin]

I tried to set the OpenVPN server on RT-AC86U (382.1_2) with custom configuration but this field is limited to 170 characters. Is it possible to make it wider? Thanks.

Not possible. Broadcom limits those variables to 255 chars, and I have to encode it in base64 to retain formatting, hence cutting it down to 170 chars.

This is a stupid HND platform limitation from Broadcom.

 

[RMerlin]

what is that all about ?

Asus has in place hardware and software-based validation checks, to prevent running Asuswrt on non-authorized hardware. Unfortunately, that check might sometimes fail if one of the radios fails to work properly. Nothing I can do about this, these checks are closed source (for obvious reasons).

 

[oldgringo]

Not possible. Broadcom limits those variables to 255 chars, and I have to encode it in base64 to retain formatting, hence cutting it down to 170 chars.

This is a stupid HND platform limitation from Broadcom.

OK, would it be possible then add more fields for openvpn options? Like push, route, topology or other specific sections from which the config would be created? Or maybe openvpnserver1.conf.add and openvpnserver2.conf.add custom config files would be an easier way?

 

[oldgringo]

Also the jffs restore fails. Once i exported it I cannot restore it back.

 

[DonCatoli]

Has anyone else found the wifi no longer working (no longer detectable) after a few hours or days with the 382 branch?, I have had to power cycle my AC3100 twice thus far since updating (rebooting does not resolve it). I will do a factory default when I have time to see if it resolves it.

 

[RMerlin]

OK, would it be possible then add more fields for openvpn options? Like push, route, topology or other specific sections from which the config would be created? Or maybe openvpnserver1.conf.add and openvpnserver2.conf.add custom config files would be an easier way?

OpenVPN already has way too many settings. Asuswrt-Merlin probably has 5x the amount of settings found on any other router that has OpenVPN support, so I'm very picky with adding any new ones.

There's already support for postconf scripts for OpenVPN.

 

[Melee]

Quite a few reports of WiFi issues after upgrading to 382 in the last few pages. My AC88 2.4GHz radio died immediately after upgrading. Factory reset and flashing back to previous firmware (including official) didn't bring it back. Had to RMA it (getting it back tomorrow). Total coincidence? Not seen such a pattern before from previous updates. Maybe ASUS themselves screwed something in the new codebase that is frying the radios?

 

[oldgringo]

OpenVPN already has way too many settings. Asuswrt-Merlin probably has 5x the amount of settings found on any other router that has OpenVPN support, so I'm very picky with adding any new ones.

There's already support for postconf scripts for OpenVPN.

Yes but many of them have almost unlimited custom config field. I use postconf atm but it's much more transparent to have all info at the same place.

 

[oldgringo]

The nvram space for client descriptions (custom_clientlist) is very limited, I can hardly set 30 devices. Is it possible to enhance the functionality similar to dhcp static list? It's weird that there are some functionalities but limited for "demo" only. It's very strange approach from Broadcom.

 

[ipowyourface]

On 382.1_2 only the front USB port on my AC88U seems to be working. It won't detect my HDD or flash drive on the rear 2.0 port. Anyone else having this issue?

 

[RMerlin]

The nvram space for client descriptions (custom_clientlist) is very limited, I can hardly set 30 devices. Is it possible to enhance the functionality similar to dhcp static list?

Not possible, since networkmap needs to be able to access that information, so I can't move it elsewhere.

 

[RMerlin]

Folks, things are what they are with the RT-AC86U. Had I known ahead of time of this nvram limitation, I would have NOT supported that model at all. So, unless Asus can come up with some solution on their end (as it also affects them), there's nothing more I can do. I can't just randomly split every nvram variable into multiple ones, as quite a few of these are accessed by Asus's closed source components, which means I cannot split these.

So, you will have to live with it. To reiterate, the same limitations also exist with the stock firmware.

 

[scjr]

Folks, things are what they are with the RT-AC86U. Had I known ahead of time of this nvram limitation, I would have NOT supported that model at all. So, unless Asus can come up with some solution on their end (as it also affects them), there's nothing more I can do. I can't just randomly split every nvram variable into multiple ones, as quite a few of these are accessed by Asus's closed source components, which means I cannot split these.

So, you will have to live with it. To reiterate, the same limitations also exist with the stock firmware.

You are working with these limitations and doing one heck of a job.

Thanks for all you do.

 

[oldgringo]

Folks, things are what they are with the RT-AC86U. Had I known ahead of time of this nvram limitation, I would have NOT supported that model at all. So, unless Asus can come up with some solution on their end (as it also affects them), there's nothing more I can do. I can't just randomly split every nvram variable into multiple ones, as quite a few of these are accessed by Asus's closed source components, which means I cannot split these.

So, you will have to live with it. To reiterate, the same limitations also exist with the stock firmware.

I understand. One of the possible solutions is push to Asus to correct these closed source components in the original firmware. But it is a long distance running.

 

[Keenan]

Folks, things are what they are with the RT-AC86U. Had I known ahead of time of this nvram limitation, I would have NOT supported that model at all. So, unless Asus can come up with some solution on their end (as it also affects them), there's nothing more I can do. I can't just randomly split every nvram variable into multiple ones, as quite a few of these are accessed by Asus's closed source components, which means I cannot split these.

So, you will have to live with it. To reiterate, the same limitations also exist with the stock firmware.

Given the limitations of the AC86U, what would be a good second choice?

 

[RMerlin]

I understand. One of the possible solutions is push to Asus to correct these closed source components in the original firmware. But it is a long distance running.

My solution is a bit hackish, I doubt they would go down the same route as me. And they already spoke with Broadcom about the upstream cause, which is from Broadcom - and no solution from Broadcom at this point.

I've seen hints of future changes Asus are working on for the RT-AC86U's nvram management, we'll see what they will head with these (sometimes they start some experimentation, then they abandon them as they don't work as expected). If they are doing what I think they are, it would probably take care of the problem once and for all on that platform.

Given the limitations of the AC86U, what would be a good second choice?

As usual, depends on your particular needs. For VPN throughput, nothing comes close to the RT-AC86U. For Wifi, the RT-AC3100 would be the closest contender.

 

[Keenan]

My solution is a bit hackish, I doubt they would go down the same route as me. And they already spoke with Broadcom about the upstream cause, which is from Broadcom - and no solution from Broadcom at this point.

I've seen hints of future changes Asus are working on for the RT-AC86U's nvram management, we'll see what they will head with these (sometimes they start some experimentation, then they abandon them as they don't work as expected). If they are doing what I think they are, it would probably take care of the problem once and for all on that platform.



As usual, depends on your particular needs. For VPN throughput, nothing comes close to the RT-AC86U. For Wifi, the RT-AC3100 would be the closest contender.

And I'm guessing as far as LAN performance the 86U would be the preferred choice due in part to its faster CPU?

 

[Vimes]

Folks, things are what they are with the RT-AC86U. Had I known ahead of time of this nvram limitation, I would have NOT supported that model at all. So, unless Asus can come up with some solution on their end (as it also affects them), there's nothing more I can do. I can't just randomly split every nvram variable into multiple ones, as quite a few of these are accessed by Asus's closed source components, which means I cannot split these.

So, you will have to live with it. To reiterate, the same limitations also exist with the stock firmware.

Without understand the technical reasons as to why you feel as you do I am glad that you have supported the 86U. Whilst it might indeed have some limitations it's VPN client performance is simply amazing, as is your modified firmware.
I'm sure that there are other commendable aspects to the 86U

 

[RMerlin]

And I'm guessing as far as LAN performance the 86U would be the preferred choice due in part to its faster CPU?

LAN performance is unaffected by the CPU. Switched Ethernet traffic will be at full gigabit regardless of the model, wireless traffic will depend on your wireless clients.

 

[Namrustler]

Does anyone have trouble with a slow VPN connection on release 382.1? I upgrade the firmware and noticed that my VPN speed drop to an average 15mb, but should be around 30 to 40mb. I can't find what's the problem is and have uploaded the attached file of the system log.

Thanks and using the RT-AC88U.

I've seen this on my RT-AC3100, but after cycling the power on it my speed over a PIA VPN to a server in my own country was 50 Mbps, which is the maximum speed of my service. Until I cycled the power, the speed was consistently between 14 and 18 Mbps.

I've only been running the new firmware for a few days and either I didn't notice an issue right away, or it occured following a subsequent reboot.