What's new

[Release] Asuswrt-Merlin 384.9 is now available

  • SNBForums Code of Conduct

    SNBForums is a community for everyone, no matter what their level of experience.

    Please be tolerant and patient of others, especially newcomers. We are all here to share and learn!

    The rules are simple: Be patient, be nice, be helpful or be gone!

I do t use smart connect and this issue is persistent. When I get some time I’ll have to give it a try again.

I'm fairly certain that is not the cause.
My devices do this and they are not mobile phones.
Some things just like to renew their connection, especially ones with cheap 2.4 GHz chips in them along with poor programming to use it.
Is it clutter? Yep, but it makes other entries standout when scrolling through.
 
I'm fairly certain that is not the cause.
My devices do this and they are not mobile phones.
Some things just like to renew their connection, especially ones with cheap 2.4 GHz chips in them along with poor programming to use it.
Is it clutter? Yep, but it makes other entries standout when scrolling through.

Or maybe you have a congested 2.4GHz environment.

I use the Wi-Fi Radar's Channel Statistics to find out the best available center channel and manually set to that channel.
 
I do t use smart connect and this issue is persistent. When I get some time I’ll have to give it a try again.

I also don't use Smart Connect.

This is what I observed. When I return from outside, my handphone will first connect to 2.4GHz, within seconds will disconnect from 2.4GHz and connect to 5GHz, then stayed on with 5GHz until I go out, which it will disconnect from 5GHz. If I set my handphone to forget the 2.4GHz password, then it will only connect/disconnect with 5GHz. There will be less log entries.
 
Apologies if this has already been covered - I've read the entire thread but may have missed something, and haven't been up to date with the rest of the forum for a couple of weeks.

I flashed 384.9 and dnsmasq wouldn't start up properly. I tracked it down to having a "dhcp-script=" entry in my /jffs/configs/dnsmasq.conf.add file. (This pointed to a script that emailed me when a new device joins my guest wifi, which I based off https://www.snbforums.com/threads/identify-guest-network-clients.37355/#post-306755.)

It seems that there is now already a "dhcp-script=/sbin/dhcpc_lease" entry in the default /etc/dnsmasq.conf, so by adding my own "dhcp-script" line to dnsmasq.conf.add I was creating a duplicate command when the .add file was merged in, and dnsmasq wouldn't start up. I don't think dhcp-script was present in /etc/dnsmasq.conf in previous releases but I could easily be wrong, will try to investigate more later.

All resolved now, though I'll have to find another way of getting those email notifications. If anyone has any insights I'd be glad to hear them, but for now I'm just posting this in case it's useful to anyone else.

PS. Almost forgot to say thanks for the update. :)

It was changed by Asus.

My recommendation would be to use a postconf script. Modify that line to call your script, and at the end of your script insert the /sbin/dhcp_lease call (make sure to also pass all arguments you received if there is any - I don't know if that's the case with dhcp-script).

I got this alsowith the last release. So I rolled back. Disappointing in Asus, again.

Why is it Asus failing if your unsupported customization (stubby) isn't compatible with their code?
 
My recommendation would be to use a postconf script. Modify that line to call your script, and at the end of your script insert the /sbin/dhcp_lease call (make sure to also pass all arguments you received if there is any - I don't know if that's the case with dhcp-script).
Thank you! Not only has that worked, but I had a "lightbulb moment" and I think the use of postconf scripts has finally clicked into place in my head. And I love learning new things even more than I like fixing problems. So thanks again.

dhcp-script receives up to 4 arguments, so can someone kindly confirm please if this would be correct as the last line of my script, in order to pass the same 4 arguments? (I'm far better at changing other people's code than I am at writing it.)
Code:
/sbin/dhcp_lease $1 $2 $3 $4
Do I need quotes? Can I just use $@ (or "$@")?
 
Last edited:
Thank you! Not only has that worked, but I had a "lightbulb moment" and I think the use of postconf scripts has finally clicked into place in my head. And I love learning new things even more than I like fixing problems. So thanks again.

dhcp-script receives up to 4 arguments, so can someone kindly confirm please if this would be correct as the last line of my script, in order to pass the same 4 arguments? (I'm far better at changing other people's code than I am at writing it.)
Code:
/sbin/dhcp_lease $1 $2 $3 $4
Do I need quotes? Can I just use $@ (or "$@")?
I do not see a pre-existing dhcp-script line in /etc/dnsmasq.conf nor a lease script in /sbin which I presume is actually your creation.
You would not put arguments to your script in the postconf script line because that is merely where the script name is set, not where it is called to run. It is called later by dnsmasq when DHCP events occur with four arguments.
This minimalist implementation is working:
Code:
# cat dnsmasq.postconf
#!/bin/sh
CONFIG=$1
source /usr/sbin/helper.sh
pc_delete "no-negcache" $CONFIG
pc_append "neg-ttl=3600" $CONFIG
pc_append "dhcp-script=/jffs/scripts/log-dhcp.sh" $CONFIG
#
. /opt/share/diversion/file/post-conf.div # Added by Diversion

# cat log-dhcp.sh
#!/bin/sh
myscriptname=`/usr/bin/basename $0`
/usr/bin/logger -t ${myscriptname} -p local6.info "Action $1, MAC $2, IP_address $3, Hostname $4"
 
Last edited:
7 days uptime and WiFi is rock solid. I have four iOS devices, two TVs and a laptop on the network. Thank you, Eric!

6 days 18 hours 23 minute(s) 34 seconds on RT-AC86U. No issues at all other than not using AiProtection...
 
Last edited:
Or maybe you have a congested 2.4GHz environment.

I use the Wi-Fi Radar's Channel Statistics to find out the best available center channel and manually set to that channel.

I'm going to guess that your theory is not applicable if Smart Connect is off like mine is.
 
6 days 18 hours 23 minute(s) 34 seconds on RT-AC86U. No issues at all other than no using AiProtection...

Is it really needed to put the minutes and seconds ?? Almost like you anticipate a failure. :rolleyes:o_O
 
Do I need quotes? Can I just use $@ (or "$@")?
Best to test it if you're not sure which will work.
a.sh
Code:
#!/bin/sh

echo '$@'
. b.sh $@
echo '"$@"'
. b.sh "$@"
echo '$*'
. b.sh $*
echo '"$*"'
. b.sh "$*"

b.sh
Code:
#!/bin/sh

echo "$# arguments"
for VAR in "$@"
do
   echo -e "\t$VAR"
done

sh a.sh "foo bar" foo bar
Code:
$@
4 arguments
        foo
        bar
        foo
        bar
"$@"
3 arguments
        foo bar
        foo
        bar
$*
4 arguments
        foo
        bar
        foo
        bar
"$*"
1 arguments
        foo bar foo bar

So "$@" would be best.
 
6 days 18 hours 23 minute(s) 34 seconds on RT-AC86U. No issues at all other than no using AiProtection...

Hmm... AiProtection works fine for me. I cleared the Malicious Sites Blocking and IPS hit count to 0 after the upgrade. Infected Device Prevention and Blocking has always been 0 all the while.

Let me recall... at the first 2 hours, the Malicious Sites Blocking and IPS hit count remain at 0 although I purposely went and visit some torrent download sites which used to produce hits.
Running "ps" in ssh terminal can see process "dcd -i 3600 -p 43200 -b -d /tmp/bwdpi/" is up and running.

Then I saw hit counts in Malicious Sites Blocking and IPS shown in Asus Router app when I accessed the router using the app. When I access using web browser again, the hit counts also shown there now.

Maybe the Asus Router app has triggered something, or maybe just coincident. But AiProtection is working fine now, including IPS hit counts.
 
my ac86u internet get disconnected when I add second openvpn client.
It's fine with the 1 openvpn client setup.
anyone having this issue?

I run a 2 VPN client setup. Previously I had upgraded to the this latest version and had experienced (what appears to be) the same issue you have - amongst others. I had tried the other suggested fixes, including switching on dns monitoring of internet connection and also disabling browser redirect notice , both of which didn't work. So, I downgraded to 384.8_2 and all is fine again. I have since been monitoring the thread.

It never occurred to me that having 2 VPN clients could cause this issue. I will upgrade again, drop 1 client (as its only there configured as a sort of hot backup for me to switch to or use) and see what happens. Thanks for the post - keep them coming!
 
my ac86u internet get disconnected when I add second openvpn client.
It's fine with the 1 openvpn client setup.
anyone having this issue?

Non-Sense... I have three (03) VPN clients connected in my AC86U and the WAN connection to the Internet is solid.
upload_2019-2-10_9-52-12.png
 
My Asus RT-AX88U keeps spewing out these messages in the system log;

Feb 10 08:07:34 kernel: blog_death_by_timeout: bad timeout value=18446744073709541616, extra_jiffies=30000, idle_jiffies=40000


It's doing it nonstop. Can't seem to find much information about it. There is clearly something amiss with my RT-AX88U as my WAN-connection has become very unstable, the WebUI will claim disconnected at random at times where it's not disconnected. But other times it might claim connected when the WAN-connection is not working. Only a reboot seems to be fixing it.

I have a Palo Alto PA-220 in parallel with it's own public IP for my Site-to-Site IPsec Connections where I route my VPN-traffic and it's not dropping out at all and as it's using the same WAN-link it does seem like it's only the RT-AX88U that is playing games with me and not the connection itself.
 
I do not think you need to specify the arguments. This minimalist implementation is working:
Thanks. It looks like there's a difference in our implementations, though. It looks like you're appending to dnsmasq.conf, which presumably means you don't already have a "dhcp-script=" line in it. My dnsmasq.conf already contains the line "dhcp-script=/sbin/dhcp_lease" so I'm using the post-conf to replace that with a call to my custom script, then calling the original "/sbin/dhcp_lease" at the end of my custom script. Possibly a difference between the two models of router?

Best to test it if you're not sure which will work.
So "$@" would be best.
Thanks! Tried those scripts here to make sure I got it. Again with the learning. (It's actually the opposite of what I would have expected, if pressed I'd have said wrapping it in quotes would concatenate the multiple arguments into one string.)
 
Dirty upgrade to 384.9 from 384.8_2. Working great so far.

Sent from my Moto G (5) Plus using Tapatalk
 
network map list is indeed unreliable
i see 8 clients connected, while i am sure more than 20 clients are
how to obtain that list of connected clients?
Me too on an AC87U. 35 clients in 384.7_2 and 11 in 384.9. All clients are connected but most that appear are being erroneously reported as wired rather than WiFi. I’m suspicious that the new firmware (the ASUS binary blob bit and networkmap specifically) may be more sensitive to weird hostnames. I have one device (a Surepet catflap) where the host name is “*” which is a likely target. I’m going to try turning all clients off and gradually reintroducing. However, before I start, immediately after the factory reset and nvram_restore/jffs_restore, I see two new clients in the list that weren’t there before. One called Asus with MAC of the 5G WiFi adapter and a 169.254 address and the other called FF:FF:FF:FF and a local IP. This address is configured by a custom startup script to give me a broadcast address for WOL while using VPN. Neither Asus nor FF appeared in the list of connected clients in the previous firmware so I fear regression may be the only option.
 
Me too on an AC87U. 35 clients in 384.7_2 and 11 in 384.9. All clients are connected but most that appear are being erroneously reported as wired rather than WiFi. I’m suspicious that the new firmware (the ASUS binary blob bit and networkmap specifically) may be more sensitive to weird hostnames. I have one device (a Surepet catflap) where the host name is “*” which is a likely target. I’m going to try turning all clients off and gradually reintroducing. However, before I start, immediately after the factory reset and nvram_restore/jffs_restore, I see two new clients in the list that weren’t there before. One called Asus with MAC of the 5G WiFi adapter and a 169.254 address and the other called FF:FF:FF:FF and a local IP. This address is configured by a custom startup script to give me a broadcast address for WOL while using VPN. Neither Asus nor FF appeared in the list of connected clients in the previous firmware so I fear regression may be the only option.

First post ( and the release notes ) :

There is a number of known issues in this release that cannot be fixed at this time:


  • Networkmap listing may be unreliable. (Bug in Asus's code)
  • Users failing to read changelogs will probably complain about the above issues. (Outside of my control).
 

Latest threads

Support SNBForums w/ Amazon

If you'd like to support SNBForums, just use this link and buy anything on Amazon. Thanks!

Sign Up For SNBForums Daily Digest

Get an update of what's new every day delivered to your mailbox. Sign up here!
Top