[Beta] Asuswrt-Merlin 380.65 Beta is now available

[bbunge]

Beta 2 working well on AC66U_B1. Even Download Master is behaving itself....

 

[Joel Teixeira]

Hello all,


On my AC3200 right after upgrade from stable (380.64_2) to 380_65_beta2 I got an exclamation mark saying the nvram was running out of space and recommending a factory reset. It was around 62k of 64k. I did as recommended but after the reset it takes already 55k. I just configured OpenVPN and MiniDLNA and I'm already on "58710 of 65536 bytes".

Is it expected 55k/64k usage on a clean reset? I'm afraid I'll reach the alarm threshold soon since I'm close already.

Is this alarm something net to 380_65 or the upgrade included new variables that used more space from my nvram?

Also, before reset I executed "nvram show | awk '{print length(), $0 | "sort -n -r"}' | cut -d"=" -f 1 | head -n 20" with the following results. I don't have a base parameter about how many entries is or reasonable, so if you could please point me some anomaly I really appreciate.

size: 63128 bytes (2408 left)
1089 sshd_hostkey
624 sshd_dsskey
547 client_info_tmp
398 rc_support
352 custom_clientlist
338 sshd_ecdsakey
335 dhcp_staticlist
120 qos_rulelist
103 dnsfilter_rulelist
96 sr_rulelist
92 2: pa5ga2
92 2: pa5ga1
92 2: pa5ga0
92 0: pa5ga2
92 0: pa5ga1
92 0: pa5ga0
86 vts_rulelist
86 asus_device_list
73 bwdpi_app_rulelist
71 vpn_server2_ncp_ciphers

 

[Ryu Connor]

RT-AC88U - 380.65 Beta1 - Reset to defaults was done

After 8 days of heavy use, the 5GHz Wi-Fi driver blew up.

Had to connect with a wired client, go to the Professional tab, and disable and then re-enable the 5GHz radio to restore communications.

The logs did not register/notice the fact that Wi-Fi connectivity on the 5GHz band died.

This is the first time I've ever seen the driver on the 88U blow up like this.

-

I've also confirmed there is a bug with the IPv6 code in the 88U in Router mode. The router will not keep a SLAAC address assigned to clients. After say an hour of steady use of the IPv6 address (say using WoW for example), the address is suddenly lost. The router will randomly re-assign the IPv6 address to the client at a later point.

Browsing IPv6 content can make this happen too (YouTube), but it takes a while, and you won't notice when the breakage happens. The browser simply just falls back to IPv4. You'd only notice by doing an ipconfig and noticing you no longer have a public IPv6 address.

I had been blaming RT-AC88U bridge mode for this bug, but I was wrong. This bug did not exist prior to the signed firmware change.

 

[RMerlin]

Is it expected 55k/64k usage on a clean reset?

On the RT-AC3200 yes. That router is starved for nvram because of the presence of three radios.

 

[bobiii]

Reporting on my rt5300. Now using beta 2 and no problems what-so-ever! All my 2.4 ghz tablets working great! ipv6 also working!

 

[ankhazam]

Hello RMerlin,
I've been toying with my AC87U on your latest firmware and its IPV6. I'm using native dhcp-pd ipv6 over PPP. The problem is to make it work I need to log in via SSH and add:

ip -6 route add default dev ppp0 metric 256

For IPV6 to be reachable by my clients.

Is it possible to add it to some kind of post PPPOE init script by so that it is done by default? This route is missing without the command and WAN IPv6 Gateway in status is empty. After adding the route it changes to: "::" and I get native ipv6 support

Sorry for the duplication but could someone else try it or copy here their rules (should they be using pppoe with native ipv6). With above route Windows machines start working like charm with ipv6. I'm still debugging Android's SLAAC sloppy network after enabling such ipv6.

Cheers,
Ankh

 

[bayern1975]

any chance to edit iptv vlan code? I need to get different vlan id instead wan vlan id?

 

[RMerlin]

any chance to edit iptv vlan code?

No.

 

[bayern1975]

No.

strange, you can edit and change everything in firmware...

 

[eddiez]

strange, you can edit and change everything in firmware...

No he (and you) can't (closed source components) and in some case does not want to, taking into account the massive impact some changes may have in other pieces of codes (or relationship with closed source parts)....

 

[bradbort]

I found all of these issues when I installed the official Asus firmware 4180 . . . Merlin merged in 4180 with beta 1 I believe which explains it. It seems I will be skipping this release until Asus fixes some of these bugs.

Me too. Too many issues with scanners and sonos on this ASUS code. I've logged an official complaint with ASUS, but who knows where that will lead....

 

[Rai80]

Sorry for the duplication but could someone else try it or copy here their rules (should they be using pppoe with native ipv6). With above route Windows machines start working like charm with ipv6. I'm still debugging Android's SLAAC sloppy network after enabling such ipv6.

Cheers,
Ankh

Same problem here! Btw, took me some time to figure out

 

[ankhazam]

Same problem here! Btw, took me some time to figure out

Yeah... for the time being I switched from SLAAC (stateless) to DHCPv6 (Stateful) native ipv6 so that my Windows PCs go with dual stack 6 & 4 whilst androids stick to ipv4.
I'm still however not sure that I should have almost 30 routes for ipv6 and whether the default one over ppp0 interface is correct -.-.

 

[triple]

380.65_beta2 works well with AC-RT87U. Thanks!

 

[visortgw]

Would it be possible for you to lessen the frequency of the DNS polling to dns.msftncsi.com? The following are recent stats from AB-Solution:

Router 1 (next most frequent host just over 400):

The top 10 requested hosts were:
--------------------------------------------------
79888 dns.msftncsi.com​

Router 2 (next most frequent host just over 2000 (i.e., trendmicro)):

The top 10 requested hosts were:
--------------------------------------------------
44918 dns.msftncsi.com​

​

 

[eddiez]

Would it be possible for you to lessen the frequency of the DNS polling to dns.msftncsi.com? The following are recent stats from AB-Solution:

Router 1 (next most frequent host just over 400):

The top 10 requested hosts were:
--------------------------------------------------
79888 dns.msftncsi.com​

Router 2 (next most frequent host just over 2000 (i.e., trendmicro)):

The top 10 requested hosts were:
--------------------------------------------------
44918 dns.msftncsi.com​

​

Would this work?

nvram set dns_probe_content=0.0.0.0
nvram set dns_probe_host=""
nvram commit
reboot


Verstuurd vanaf mijn SM-G935F met Tapatalk

 

[RMerlin]

Would this work?

nvram set dns_probe_content=0.0.0.0
nvram set dns_probe_host=""
nvram commit
reboot


Verstuurd vanaf mijn SM-G935F met Tapatalk

The proper way to disable it is by clearing dns_probe_content. Wanduck will then consider the probe as being disabled.

nvram set dns_probe_content=""
nvram commit

No idea what impact it will have on the router's behaviour however.

 

[RMerlin]

Would it be possible for you to lessen the frequency of the DNS polling to dns.msftncsi.com?

Good luck figuring out how the wanduck code works. I gave up a long time ago - it's a complete mess...

 

[o-l-a-v]

Would it be an idea, if its possible and not too much of a hassle, to add a checkbox next to the factory reset button, that would clear nvram at the same time? Would be pretty neat.

 

[thelonelycoder]

Would it be possible for you to lessen the frequency of the DNS polling to dns.msftncsi.com? The following are recent stats from AB-Solution:

The proper way to disable it is by clearing dns_probe_content. Wanduck will then consider the probe as being disabled.

AB-Solution 3.1 will have an 'experimental settings' menu.
Among the settings options is the option to delete or restore the dns_probe_content in nvram.
It bothers me to no end seeing these queries every ten seconds in the dnsmasq.log file.