Asuswrt-Merlin 378.50 Beta 2 is out

[RMerlin]

Regarding the problems with vsftp, from what I recall OpenWRT folks had some issues when the upgrade was done there too. I actually found these as well:

the bug fix to move from 2.x to 3.0.0:
https://dev.openwrt.org/changeset/31241

later path fixes:
https://github.com/openwrt/packages...8fd5203#diff-9c9c3a2f0d7a62d2e4bba810b6cc3312

I had checked the later patches from Github, but hadn't looked at the older commits from their svn repo. That one doesn't seem related to stability however, but to compile issues, as per this ticket:

https://dev.openwrt.org/ticket/11263

 

[RMerlin]

Since I already made the code intelligent enough not to feed dnsmasq with an invalid hostname from the DHCP static list, I decided to link the name entered on the networkmap with the DHCP static list. Any invalid hostname will simply not be used by dnsmasq when providing static leases.

 

[Speedy1205]

Are you able to enable 80mhz channel bandwidth on 5ghz band on the AC56U + Beta 2 DPI build?

Yes I`m actual using 80mhz with Auto Channel Selections. It works fine and is stable for me.

 

[RMerlin]

I didn't get any reaction to my post regarding this issue in thread 378.50 Beta1 and I am trying again.
With release 378.50 on my RT-AC87U I have noticed the following with the generated /tmp/resolv.dnsmasq: when I enter No in "Connect to DNS Server automatically" in WAN DNS Setting, the value 8.8.8.8 for "DNS Server1" and click Apply the software generates a resolv.dnsmasq in /tmp containing:
server=/local/8.8.8.8

Forward local domain queries to upstream DNS is set to No in DHCP Server, DNS and WINS Server Setting.

By comparaison version 376.49_50 \tmp\resolv.dnsmasq is empty, not trace of server=/local/8.8.8.8

I don't know what are the exact consequences of that change. I would like to avoid all my DNS queries for the local (local=lan) domain to be forwarded to upstreams DNS servers. Could somebody help me out ?

AFAIK, that does not mean "forward any local domain", it only means "forward THE domain called local". I.e. for lookups on computer.local or tablet.local.

Why Asus is specifically forwarding anything ".local" upstream, I don't know - you'd have to ask them. They did that change in recent updates.

Enabling the option to not forward the local domain upstream will add an additional line to your config file specifying what your local domain is, and to only use DHCP/host file to resolve anything within that specific domain.

 

[cyborgpunte]

Confirmed I must be stupid I don't understand this.
Why two occurrence ???
please help me out
(rt-ac68u, 1.0.2.1, 378.50_beta2 and tried clean reflash + factory defaults and deleted syslog.log)

#!/bin/sh
logger "mounting CIFS drive in post-mount 	why"
mkdir /tmp/mnt/solo2
mount \\\\172.16.16.209\\Filmer /tmp/mnt/solo2 -t cifs  -o "username=,password="

result in

Dec  1 01:00:18 admin: mounting CIFS drive in post-mount 	why
.......
Dec  1 01:00:44 admin: mounting CIFS drive in post-mount 	why

 

[john9527]

Confirmed I must be stupid I don't understand this.
Why two occurrence ???
please help me out

Do you have more than 1 drive attached? I think post-mount gets called for each drive that is is mounted.

 

[cyborgpunte]

Thanks @john9527 you are right.

Detached one of my physical drives and all of the sudden just one occurrence
So perhaps it was just a bit of bad luck for me when I was thinking

What is the latest executed jffs-script ??

 

[john9527]

I'm pretty sure that the name of the mount point is passed as an argument to the post-mount script, so you could check that to only execute your commands when the proper drive is mounted,

 

[NGRhodes]

AC56U.
Upgraded from DD-WRT, using the CFE recovery console to wipe and install.
WiFi, Guest access, UI all seemed fine.

Adaptive QoS seems not to work properly though - 99% of my test traffic (eg. speedtest.net and torrents) get put in general category.
Setting specific machines to high and low made no difference to which machine got priority of traffic.

 

[Martineau]

Detached one of my physical drives and all of the sudden just one occurrence
So perhaps it was just a bit of bad luck for me when I was thinking

What is the latest executed jffs-script ??

If the question is "which is the last jffs script to finish executing?", then the answer is "Well it depends!"

RMerlin did say that as the jffs scripts are event driven, the precise order cannot be guaranteed, however, in my setup, init-start is the last to finish executing!

The reason for this is because I wanted all my customisation messages to appear (where possible) in the same section of the syslog, so I basically set a semapahore which can be checked by all the other scripts to serialise them in a strict sequence.

Here is how my init-start script is always the first to start and the last to finish......

#!/bin/sh

HARDWARE_MODEL=$(nvram get productid)
MYROUTER=$(nvram get computer_name)
BUILDNO=$(nvram get buildno)

/usr/bin/logger -s -t "($(basename $0))" $$ "Martineau $MYROUTER BOOT in progress... [$@]" 

# NOTE: Can't use Flash drive /tmp/mnt/RT-AC56U/ 'cos it hasn't been mounted yet :-(
echo `date` > /tmp/BOOTINPROGRESS

# If we ensure that the mount point exists at this stage of the boot, it can eliminate the Hotplug boot error?
#   hotplug[427]: USB /dev/sdxn (ext4) failed to mount at the first try! 
#mkdir -p /tmp/mnt/$MYROUTER

# Should be sufficient to cover physical BOOT process?
sleep 150


logger -s -t "($(basename $0))" $$ LEDs ON scheduled 06:00
/usr/sbin/cru a LEDsON "0 6 * * * /jffs/scripts/ledson.sh"

<snip>

# Move Syslog to USB Flash drive
/jffs/scripts/syslog-move.sh

rm /tmp/BOOTINPROGRESS


/usr/bin/logger -s -t "($(basename $0))" $$ "Martineau $MYROUTER BOOT Completed Firmware build $BUILDNO [$@]"

 

[Aldebaran]

Wireless MAC filter doesn't block unknown devices

My wireless MAC filter ignores a new device (a Kobo Aura e-reader). I have a MAC list enabled on 'Allow Access' for both 5 and 2.4 Ghz. Normally I expect that a new device is unable to use the wireless network. In my case the Kobo is able to enter my wireless network en connect to the internet.

I upgraded to firmware version is 376.49_5 after I noticed the issue on V376.44.
Any ideas how this is possible?

 

[RMerlin]

My wireless MAC filter ignores a new device (a Kobo Aura e-reader). I have a MAC list enabled on 'Allow Access' for both 5 and 2.4 Ghz. Normally I expect that a new device is unable to use the wireless network. In my case the Kobo is able to enter my wireless network en connect to the internet.

I upgraded to firmware version is 376.49_5 after I noticed the issue on V376.44.
Any ideas how this is possible?

Please don't hijack threads. This thread is about gathering feedback on the 378.50 beta releases.

 

[themrindy]

AICloud + lan "issue"?

Hi,

I have used 378.50_2 (rt87u) for a week now, very stable!

I got a new version for the AICloud app to my IOS devices and now I cant browse my connected usb3 disk (fine if I go through the website). Anyone will similar issues?

and a question:
When I do a speedtest I get without the router connected about 700/700Mbit with the router connected I get 400/700Mbit. The thing I have noticed that the CPU utilizes 100% (core 1) when I test the download and only around 30-40% for the upload. The test is done at Windows 7 x64 4Gb with a 1Gbit NIC (cat6).
Could it something in that area that holds my number down?

I have NAT enabled and firewall but no QoS or traffic analyzes is enabled.

Will this router "eat" so much download speed?? I hope not...

Regards,
Christian

 

[bigmag]

@Rmerlin

Hi

Wen i set the Firewall - URL Filter of on the max Download Bandwidth is 150MB

i hade a conction of 200/10 wen i set the Firewall - URL Filter off ist always ok

 

[wgentine]

RMerlin,

Any way to have more stronger sinal on 2.4 radio for final release?

When compared to R7000, the signal is not so strong as Asus uses issues less power.

 

[M4tt]

Since I upgraded my RT-AC68U to 378.50b2 I have an issue with my 2.4Ghz Wifi:

If I set "Wireless Mode" to "Auto", my Macbook Pro (Late 2013, Yosemite 10.10.2) only connects with 11MBit/s. When I set it to "N only", I get 217 MBit/s.

I am sure this is not related to my settings. When I do a factory reset, this issue is present just after setting up user/pw,wifi name and wpa2 pw

 

[RogerSC]

Those pretty much look like the standard firewall rules, so everything looks ok from the router's end.

You'll have to see what explanation that test site gives for those three specific tests it fails to determine what's going on. Personally, I never blindly trust such test sites.

You know, this is really odd. If I run this test right after flashing or turning on the router, it gets 17/20. If I wait about half an hour or so, and don't do anything to the router, that same score goes up to 19/20. Don't know why that would be, should be the same rules, just a half-hour older *smile*.

The difference between 17/20 and 19/20 is whether or not ICMPv6 packets coming to me are filtered. If they're filtered, I see the lower score, if they're not filtered I see the higher score.

Like I said, weird results. So I believe you can disregard this report functionally, but it is maybe food for a minute's thought *smile*.

 

[john9527]

You know, this is really odd. If I run this test right after flashing or turning on the router, it gets 17/20. If I wait about half an hour or so, and don't do anything to the router, that same score goes up to 19/20. Don't know why that would be, should be the same rules, just a half-hour older *smile*.

The difference between 17/20 and 19/20 is whether or not ICMPv6 packets coming to me are filtered. If they're filtered, I see the lower score, if they're not filtered I see the higher score.

Like I said, weird results. So I believe you can disregard this report functionally, but it is maybe food for a minute's thought *smile*.

I believe the IPv6 spec does not allow blocking echo requests, so that's the difference. You may want to double check your host firewall settings to see if it is doing the blocking. For example, in another thread someone noted that Kapersky default was to block IPv6 echo.

 

[RMerlin]

I believe the IPv6 spec does not allow blocking echo requests, so that's the difference. You may want to double check your host firewall settings to see if it is doing the blocking. For example, in another thread someone noted that Kapersky default was to block IPv6 echo.

Right. Also, let's not forget that with IPv6, you have a whole subnet routed to you, it's not one single IP that's behind NAT. So, that failure might be with the computer from which you are testing rather than with the router (which simply routes that traffic to your client).

 

[pablo11]

Overclock button

I know I am probably stupid to ask this but - why does my boot-loader (1.0.2.0) say that my router is not able to overclock and yet it is a feature of your firmware?

I must say thanks very much because everything else is absolutely fine

If I do press the overclock button my 5ghz band is unusable....

I am on RT-AC68U Australia.