[Release] Asuswrt-Merlin 384.11 is available

[John Adler]

I am happy to see again all the devices connected to my network!
since 384-08 I can see them again

 

[JGrana]

With folks having issues with restoring settings after a factory reset, here is a script that I had used in the past to at least cushion the longest part of restoring - redoing DHCP static assignments...
It also gets around the nvram size limitation in more recent ASUS firmwares (where I got "bit" ;-)

It reads the nvram dhcp_staticlist, sorts the names, then outputs a line for each host that is compatible with dnsmasq, something like this:

dhcp-host=00:01:48:FF:31:C1,192.168.1.121,TestH1

Here is the script. You can copy/paste into a file. Frankly it really only needs to run once since going forward you will only update the dnsmasq.conf.add file.

nvram get dhcp_staticlist | sed 's/</\n/g' | grep ":" | awk -F">" '{ print "<"$2">"$1">"$3; }' | \
sort -t . -k 1,1n -k 2,2n -k 3,3n -k 4,4n | awk -F">" '{ print "dhcp-host="$2","$1","$3; }'| sed s/\<//

What I had done is run the above script and re-direct the output to a file called dnsmasq.conf.add.
You can then clear the nvram dhcp_staticlist and/or disable Manual Assignment in the GUI.
Put a copy of dnsmasq.conf.add in /jffs/configs.
Reboot and dnsmasq will now use this list to assign fixed IP addresses. You can add, delete, change names, etc. to /jffs/configs/dnsmasq.conf.add. If you do make any changes, just restart dnsmasq.conf

service restart_dnsmasq

I always store 2 copies - one in /jffs/configs and another for updating, changing and for backup up (on an external USB stick).

 

[skeal]

With folks having issues with restoring settings after a factory reset, here is a script that I had used in the past to at least cushion the longest part of restoring - redoing DHCP static assignments...
It also gets around the nvram size limitation in more recent ASUS firmwares (where I got "bit" ;-)

It reads the nvram dhcp_staticlist, sorts the names, then outputs a line for each host that is compatible with dnsmasq, something like this:

dhcp-host=00:01:48:FF:31:C1,192.168.1.121,TestH1

Here is the script. You can copy/paste into a file. Frankly it really only needs to run once since going forward you will only update the dnsmasq.conf.add file.

nvram get dhcp_staticlist | sed 's/</\n/g' | grep ":" | awk -F">" '{ print "<"$2">"$1">"$3; }' | \
sort -t . -k 1,1n -k 2,2n -k 3,3n -k 4,4n | awk -F">" '{ print "dhcp-host="$2","$1","$3; }'| sed s/\<//

What I had done is run the above script and re-direct the output to a file called dnsmasq.conf.add.
You can then clear the nvram dhcp_staticlist and/or disable Manual Assignment in the GUI.
Put a copy of dnsmasq.conf.add in /jffs/configs.
Reboot and dnsmasq will now use this list to assign fixed IP addresses. You can add, delete, change names, etc. to /jffs/configs/dnsmasq.conf.add. If you do make any changes, just restart dnsmasq.conf

service restart_dnsmasq

I always store 2 copies - one in /jffs/configs and another for updating, changing and for backup up (on an external USB stick).

This is awesome!

 

[SheikhSheikha]

Did you test with Auto or manual Control Channels? How far apart are the routers and what is in-between them?

Maybe posting a few of your wireless setup pages will help others give further suggestions?

See the 2.4GHz settings of my router and the network map of the repeater. The connection is not the issue, neither the distance. The distance is approximately 10 meters between router and repeater AC5300. The repeater runs smoothly and fast at 5GHz. The 2.4GHz on the repeater connects but no internet is reached when connected to it.
The router runs smooth on 5GHz, but the 2.4GHz drops after a while from 600mbps to 150mbps.
The 40GHz settings on 2.4GHz worked fine until the latest firmware. Before no issues with it.
NVRAM was erased, M&M was done. JFFS was erased. No scripts No Qos no AiProtect.
The router is connected by wire to an ISP router/modem (some TP-Link shirt) that is mandatory. The AC5300 router does everything (DNS, ExpressVPNClient in L2TP, firewall, iP6 is off. Nothing fancy. Straight router. USB is on 2.0. Any help is welcome and appreciated to solve this

 

[NGRhodes]

Upgraded yesterday.
Noticed reported CPU (reported by Network Map screen) is about half previously when running speedtests over wifi.
SIRQ is staying in mid 20% according to top (75MB down connection) , unfortunately never took note of what it was before. CTF is on (as it was before upgrading). I am running Adaptive QoS over PPPoE connection.
Oh and my network map is working again, even picking up correct machines names (where before some names were mac addresses).
Potential to be a very very solid release, thanks !

 

[elgreco29]

The nvram save/restore utility has not been compatible with Merlin’s firmware for a good while now (since at least the 384 series and possibly the 382 firmware series. (But it still works with John’s fork.). If it worked for you, you’ve been really lucky, but at the first sign of any problems, get ready to do a reset with a manual input of your settings. And if you follow L&LD’s Guide, to the letter, you’ll be fine.

https://www.snbforums.com/threads/n...l-and-manual-configuration.27115/#post-205573

Well worth reading the whole thread.

Thanks for the hint. Until today i had no problems with this utility. Is there any reason why this helpful tool is not supported anymore?

 

[JemTheWire]

Now all we need is for JackYaz to add it to scMerlin [emoji3]

 

[HawkInOz]

The issue only affects the Network Analysis page under Network Tools, don't let that stop you unless for some reason who need to rely on this page on a regular basis. It's just two build time options related to Network Tools that didn't get properly disabled in the merge. Everything else is fine.

So, no new build for the affected models until the next build for all of the supported models (such as 384.11_2, should it be necessary)? Thanks.

 

[joe scian]

See the 2.4GHz settings of my router and the network map of the repeater. The connection is not the issue, neither the distance. The distance is approximately 10 meters between router and repeater AC5300. The repeater runs smoothly and fast at 5GHz. The 2.4GHz on the repeater connects but no internet is reached when connected to it.
The router runs smooth on 5GHz, but the 2.4GHz drops after a while from 600mbps to 150mbps.
The 40GHz settings on 2.4GHz worked fine until the latest firmware. Before no issues with it.
NVRAM was erased, M&M was done. JFFS was erased. No scripts No Qos no AiProtect.
The router is connected by wire to an ISP router/modem (some TP-Link shirt) that is mandatory. The AC5300 router does everything (DNS, ExpressVPNClient in L2TP, firewall, iP6 is off. Nothing fancy. Straight router. USB is on 2.0. Any help is welcome and appreciated to solve this

Set channel BW to 20 Mhz for stable connection. Protected mgmt frames to capable. Bluetooth coex - disable. Preamble - short
Tx bursting - disable. Optimize ampdu - enable

 

[Xentrk]

Check what your WAN interface is, in case it's not eth0:

nvram get wan0_ifname

Also make sure you don't use a VPN, which would route all traffic through it.

I recalled seeing that posted earlier and did the nvram command to validate the interface. It does return eth0. I do have Accept DNS Configuration = Disabled in the OpenVPN Client. I also tried shutting down the OpenVPN client when running the command to take the VPN out of the mix. I will do another review of settings to see if I can find out what may be the issue.

Update
Changing to ppp0 gets me output:

tcpdump -i ppp0 -p port 853 -n

01:43:16.513189 IP 1.1.1.1.853 > x.x.x.x.45266: Flags [P.], seq 3964:4372, ack 811, win 31, length 408
01:43:16.513484 IP x.x.x.x.45266 > 1.1.1.1.853: Flags [.], ack 4372, win 669, length 0
01:43:16.685148 IP 1.1.1.1.853 > x.x.x.x.45266: Flags [P.], seq 4372:5292, ack 811, win 31, length 920
01:43:16.715715 IP x.x.x.x.45266 > 1.1.1.1.853: Flags [.], ack 5292, win 760, length 0

 

[Xentrk]

With folks having issues with restoring settings after a factory reset, here is a script that I had used in the past to at least cushion the longest part of restoring - redoing DHCP static assignments...
It also gets around the nvram size limitation in more recent ASUS firmwares (where I got "bit" ;-)

It reads the nvram dhcp_staticlist, sorts the names, then outputs a line for each host that is compatible with dnsmasq, something like this:

dhcp-host=00:01:48:FF:31:C1,192.168.1.121,TestH1

Here is the script. You can copy/paste into a file. Frankly it really only needs to run once since going forward you will only update the dnsmasq.conf.add file.

nvram get dhcp_staticlist | sed 's/</\n/g' | grep ":" | awk -F">" '{ print "<"$2">"$1">"$3; }' | \
sort -t . -k 1,1n -k 2,2n -k 3,3n -k 4,4n | awk -F">" '{ print "dhcp-host="$2","$1","$3; }'| sed s/\<//

What I had done is run the above script and re-direct the output to a file called dnsmasq.conf.add.
You can then clear the nvram dhcp_staticlist and/or disable Manual Assignment in the GUI.
Put a copy of dnsmasq.conf.add in /jffs/configs.
Reboot and dnsmasq will now use this list to assign fixed IP addresses. You can add, delete, change names, etc. to /jffs/configs/dnsmasq.conf.add. If you do make any changes, just restart dnsmasq.conf

service restart_dnsmasq

I always store 2 copies - one in /jffs/configs and another for updating, changing and for backup up (on an external USB stick).

Sweet. I do mention the technique to use dnsmasq.conf.add for dhcp static ip reservations in my blog post https://x3mtek.com/asuswrt-merlin-firmware-upgrade/. Your small script helps with the conversion. Do you mind if I update the blog post with your script? I would also like to add it to my repo for miscellaneous scripts on https://github.com/Xentrk/Asuswrt-Merlin-Linux-Shell-Scripts. I will provide the necessary acknowledgement/credit to you.

A modification to append the lines to dnsmasq.conf.add would be a nice addition.

I also wrote a script to save and restore static DHCP reservations that I can add to my miscellaneous library.

 

[RMerlin]

So, no new build for the affected models until the next build for all of the supported models (such as 384.11_2, should it be necessary)? Thanks.

There will be new releases, however I am not in a hurry since these are not critical issues.

 

[martinr]

Thanks for the hint. Until today i had no problems with this utility. Is there any reason why this helpful tool is not supported anymore?

From the author himself


“The big problem with using the tool to migrate from 380 up to 382/384 is the new length limits imposed on the nvram vars (there's a separate allowed length parameter specified for each variable). If you have a variable that exceeds the limits, dhcp_staticlist for example for someone with a lot of manual assignments) it will silently fail the restore.”

https://www.snbforums.com/threads/backup-manual-dhcp-list.12876/page-4#post-384376

 

[WMConey]

No idea then. The networkmap data is generated by a closed source component. The name is typically extrapolated by it based on various factors, such as the DPI engine's identification of the client, Netbios name (if any), etc...

Maybe try removing /jffs/nmp_cl_json.js through SSH then reboot (also through SSH) and see what happens.

Something has changed in the way Network map generates default names. For example, I suddenly [after upgrade] had a DCHP attached device that was listed as "IEEE Certification Authority." As you might expect, this both bemused and troubled me - turned out to be a WiFi-enabled electrical outlet I had recently added to my SmartHome setup. No idea how this was picked as the identifier.

 

[RMerlin]

Something has changed in the way Network map generates default names. For example, I suddenly [after upgrade] had a DCHP attached device that was listed as "IEEE Certification Authority." As you might expect, this both bemused and troubled me - turned out to be a WiFi-enabled electrical outlet I had recently added to my SmartHome setup. No idea how this was picked as the identifier.

I know networkmap now does OUI lookups to help identifying clients. So if you have a custom/weird MAC address, it may affect results.

 

[SheikhSheikha]

Set channel BW to 20 Mhz for stable connection. Protected mgmt frames to capable. Bluetooth coex - disable. Preamble - short
Tx bursting - disable. Optimize ampdu - enable

Thank you. Will give that a try and revert back with my experiences.

 

[Protik]

Upgraded to 384.11 without any issues on my AC68U. But my System log is flooeded with the following,

May 11 23:09:17 acsd: scan in progress ...
May 11 23:09:17 acsd: scan in progress ...
May 11 23:09:18 acsd: scan in progress ...
May 11 23:09:18 acsd: scan in progress ...
May 11 23:09:18 acsd: scan in progress ...
May 11 23:09:18 acsd: scan in progress ...
May 11 23:09:19 acsd: scan in progress ...
May 11 23:09:19 acsd: scan in progress ...
May 11 23:09:19 acsd: selected channel spec: 0xe29b (157/80)
May 11 23:09:19 acsd: Adjusted channel spec: 0xe29b (157/80)
May 11 23:09:19 acsd: selected channel spec: 0xe29b (157/80)

And,

May 12 05:52:22 netdata[928]: clock_gettime(7, ×pec) failed.
May 12 05:52:22 netdata[928]: clock_gettime(7, ×pec) failed.
May 12 05:52:23 netdata[928]: clock_gettime(7, ×pec) failed.
May 12 05:52:23 netdata[928]: clock_gettime(7, ×pec) failed.
May 12 05:52:23 netdata[928]: clock_gettime(7, ×pec) failed.
May 12 05:52:23 netdata[928]: clock_gettime(7, ×pec) failed.
May 12 05:52:24 netdata[928]: clock_gettime(7, ×pec) failed.
May 12 05:52:24 netdata[928]: clock_gettime(7, ×pec) failed.
May 12 05:52:24 netdata[928]: clock_gettime(7, ×pec) failed.
May 12 05:52:24 netdata[928]: clock_gettime(7, ×pec) failed.
May 12 05:52:25 netdata[928]: clock_gettime(7, ×pec) failed.
May 12 05:52:25 netdata[928]: clock_gettime(7, ×pec) failed.
May 12 05:52:25 netdata[928]: clock_gettime(7, ×pec) failed.
May 12 05:52:25 netdata[928]: clock_gettime(7, ×pec) failed.

Any idea what's going on here?

 

[RMerlin]

May 11 23:09:17 acsd: scan in progress ...

This is the channel selector daemon simply logging what it's doing. This is because you have your channel set to Auto.

May 12 05:52:22 netdata[928]: clock_gettime(7, ×pec) failed.

No idea what that is, there is no netdata process in the firmware, so it must be something you installed.

 

[Martineau]

Upgraded to 384.11 without any issues on my AC68U. But my System log is flooeded with the following,

May 12 05:52:22 netdata[928]: clock_gettime(7, ×pec) failed.

Any idea what's going on here?

I think this was discussed on Netdata Github

clock_gettime with clockid_t clk_id=7 is a request to get the CLOCK_BOOTTIME.

man 2 clock_gettime says that that's available after 2.6.39:

      CLOCK_BOOTTIME (since Linux 2.6.39; Linux-specific)

although it implied a kernel version issue?

I think the messages started appearing in perhaps v1.9.xxx so I used syslog-ng to filter the netdata messages, but I think they have now stopped on my current installation.

Check your version of 'netdata - 1.12.1-1a'.

 

[RamGuy]

With folks having issues with restoring settings after a factory reset, here is a script that I had used in the past to at least cushion the longest part of restoring - redoing DHCP static assignments...
It also gets around the nvram size limitation in more recent ASUS firmwares (where I got "bit" ;-)

It reads the nvram dhcp_staticlist, sorts the names, then outputs a line for each host that is compatible with dnsmasq, something like this:

dhcp-host=00:01:48:FF:31:C1,192.168.1.121,TestH1

Here is the script. You can copy/paste into a file. Frankly it really only needs to run once since going forward you will only update the dnsmasq.conf.add file.

nvram get dhcp_staticlist | sed 's/</\n/g' | grep ":" | awk -F">" '{ print "<"$2">"$1">"$3; }' | \
sort -t . -k 1,1n -k 2,2n -k 3,3n -k 4,4n | awk -F">" '{ print "dhcp-host="$2","$1","$3; }'| sed s/\<//

What I had done is run the above script and re-direct the output to a file called dnsmasq.conf.add.
You can then clear the nvram dhcp_staticlist and/or disable Manual Assignment in the GUI.
Put a copy of dnsmasq.conf.add in /jffs/configs.
Reboot and dnsmasq will now use this list to assign fixed IP addresses. You can add, delete, change names, etc. to /jffs/configs/dnsmasq.conf.add. If you do make any changes, just restart dnsmasq.conf

service restart_dnsmasq

I always store 2 copies - one in /jffs/configs and another for updating, changing and for backup up (on an external USB stick).

This sure is a great recommendation. It seems like you have some experience with using *.conf files for various configurations. Do you know if there is a way to configure static DNS entries using something like this? The only real way to get static DNS entries when using the Asus as local DNS server currently is by using static DHCP mapping as using the static list enforces whatever hostname you set in the static list to also be resolvable by the local DNS on the router.

But what if I want to have hostname resolve to several IP-addresses and in my case I would want to add static IPv6 entries in the DNS as there seems to be no way to configure static DNS for IPv6 addresses in the WebUI. Or what if I want to set static entries for IP-addresses that are not local but are behind a site-to-site VPN connection so they are still using private addresses that I want to be resolvable. Are you aware of any capabilities like this?