[Fork] Asuswrt-Merlin 374.43 LTS releases (Archive)

[GHammer]

Interesting.....as I said, I'm surprised it loaded. But just for grins.....here's a description of rdnssd.....

rdnssd auto configures recursive DNS servers on IPv6 networks using ICMPv6 Neighbor Discovery (RFC 5006), and can update the DNS resolvers configuration....

Did your get this error with the filter option enabled or disabled?

Filter is enabled, that is the default, isn't it? At least I have not changed it this build or any other.

 

[john9527]

After that, it restarts. This only happens when the 2GHz band is used as the repeater. Any clue what's going on here?

It took a while for my memory to kick in.....is this the same problem we had in the past where I provided you a test build? Could this be the same issue with the implementation of the IPv6 broadcast RFC when connected to Apple. I can do another test build, but if I remember correctly when I looked at it last time, I can't make it an nvram option because it's a change in the kernel code.

 

[john9527]

Filter is enabled, that is the default, isn't it? At least I have not changed it this build or any other.

Yes, I didn't change the default from enabled.

I'd still like to take a look at this....is there somewhere you could share your entire syslog after a boot?

 

[GHammer]

Sure. I'll PM you a link shortly.

 

[Ford Prefect]

No change in anything that would affect the wireless......are these by chance Apple clients? Seems to be hit or miss with them lately.....

...it actually startet with a Lenovo W520 Thinkpad running Win7Pro-64bit, two weeks ago, what normally had rock solid connections "since years".
But indeed, since yesterday there is an iphone 5s in the family.

...the effect has gone after i switched back to v7 of the fork.
Never had problems with Kindles (paperwhites, but these also use 2.4GHz only).

 

[Trebuin]

It took a while for my memory to kick in.....is this the same problem we had in the past where I provided you a test build? Could this be the same issue with the implementation of the IPv6 broadcast RFC when connected to Apple. I can do another test build, but if I remember correctly when I looked at it last time, I can't make it an nvram option because it's a change in the kernel code.

We worked with something Comcast IPv6 related, I think it was someone else who worked with this. I've had IPv6 disabled because it causes issues even with my router. Do you suspect this is the same thing that's happening on 5GHz, just not reported?

I'm trying to remember...you added a firewall feature that addressed the issue that I had before...might have been the same one so I'll give it a try and check in tomorrow to see how things are going with it.

Edit: It seems you may be right, referenced below. Maybe I can work around this...do you know of what I can change on my apple router to stop this issue? Internet options IPV6 on the apple router has been set to local link only.
////
- I've been trying to think of anything that could possible cause a problem in the latest build. Only thing I remotely could come up with was an
IPv6 fix that disables the anycast address for /127 and /128 prefixes per RFC, so I backed that one out for a test.

RT-AC68U_3.0.0.4_374.43_2-08B3j9527.trx
RT-AC66U_3.0.0.4_374.43_2-08B3j9527.trx
/////
Edit 2:
After finding and reviewing my old issue, it does seem to be acting the same in a way, but not nearly as bad as it did before. Before, it was quite frequently...maybe once every two or three minutes...now it's only once every 2-3 hours. I tested it before using speedtest and it would show the drop to zero, but this now is a total internet disconnect. The two routers definitely stop talking.
////
Edit 3:
On version 8's release notes-
* NEW: firewall: add forward rule for ipv6 dhcp - maybe fix Comcast connect?

It's not there anymore, is it?

 

[Builder71]

...

• • New options (Note: Fork unique options are designated by '**' in the gui)
    • Added gui option to save https generated certification (persistent across reboots)

Nice!
What is the above new option about?

 

[john9527]

Added gui option to save https generated certification (persistent across reboots)

Right now if you have https access enabled to the router, it will generate a self-signed certificate. Then when you connect to the router, your browser will flag the certificate since it's self-signed, and you can then install the certificate to your certificate store for future access without any warnings. The problem is that in the default setup, the certificate will be regenerated on every boot requiring you to re-install the certificate. This gives you the option of saving the certificate (in nvram) so it doesn't get regenerated on every boot. Then you can install the certificate once and be done with it.

 

[john9527]

Apr 9 00:40:02 rc_service: wlcconnect 339:notify_rc restart_wlcmode 0

I tried to trace back this msg and I think it's a symptom, not a cause. It looks like the router is reacting to a state change action that the repeater has been disconnected and it enters a disconnect/cleanup sequence.

Is the 2.4 GHz band particularly 'busy' in your environment where there could be interference with the AP?

 

[GHammer]

@john9527
Were you able to get the log I made available for you yesterday?

 

[john9527]

@john9527
Were you able to get the log I made available for you yesterday?

Yes, I did. Was actually one of the cleanest logs I ever looked at. Wasn't able to determine much though about the rdnssd msg. It looks like several time sensitive services are being restarted after the ntp time sync occurs, one of which is rdnssd. It then waits for ~1 sec, checks the progress, and sees rdnssd hasn't yet started, and issues that msg. Then ~2 secs later, rdnssd is up and running. So, it's not anything I'd be concerned about...it looks like a tight timing quirk.

I see you updated your signature to N66U....that make more sense. In the logs I see where the Broadcom wireless drivers are successfully loaded...which would be tough since the N56U has RAlink wireless hardware.

 

[GHammer]

I like tidy!
Yes, I have an N66U, sorry for the fat fingers.
Since I have not noticed anything amiss, I'll just tell myself that the service did start but not before the supervisor decided it aborted.

FWIW- Running fine here and I like the cleaner look of the GUI.

 

[Trebuin]

I tried to trace back this msg and I think it's a symptom, not a cause. It looks like the router is reacting to a state change action that the repeater has been disconnected and it enters a disconnect/cleanup sequence.

Is the 2.4 GHz band particularly 'busy' in your environment where there could be interference with the AP?

I think you're exactly on the spot with the 2.4GHz...probably the 5GHz if it just isn't reporting it. It acts the same. Anyhow, I have 5 clients all with -5x db or better connected to the 2.4 (now acting as the repeater as well) & 2 clients on the 5GHz. No one else is using 5GHz near me and for 2.4 GHz:
1 transmitter hard wired to the primary AP Ch11 (Channel 11)
Primary AP sees several Ch6 and Ch1 routers. I can't change the CH11 on the repeater.
Repeater sees the Ch6 and Ch1 routers and 1 Ch11 router from a neighbor, -80db signal so pretty low.
Primary AP has 1 5Ghz client and 2 2.4GHz clients.

Since I've disabled the IPV6 firewall last night, I have seen no issues on the 2.4. I'll check again tomorrow. If it remains fine, I'll reconnect to the 5GHz freq and check that.

As of now, potential issue with IPV6 causing the repeater to restart. I'll really have to play around with IPV6, it's been about 9 months since I disabled it. Let me know if you want me to try anything after waiting another day for the reliability check.

 

[CyberMew]

Got a question here. I'm using google dns servers right now, and want to use another dns server for a specific device. I'm not sure if DNS filtering (under Parental control) is the right place to do it?

Also, whenever I modify my dns servers (under WAN), it doesn't take effect immediately. it seems that it takes a long time before changes are actually made. And I tried to reboot the router and it's the same thing. It only takes effect like an hour later. Any idea what's wrong?

 

[RMerlin]

Got a question here. I'm using google dns servers right now, and want to use another dns server for a specific device. I'm not sure if DNS filtering (under Parental control) is the right place to do it?

That's what DNSFilter was designed for, yes.

Also, whenever I modify my dns servers (under WAN), it doesn't take effect immediately. it seems that it takes a long time before changes are actually made. And I tried to reboot the router and it's the same thing. It only takes effect like an hour later. Any idea what's wrong?

The DNS entered on the WAN page is what gets sent to your clients as they contact the DHCP server on your router to obtain a lease. If you want any change there to immediately apply to your clients, you need to restart your clients, not the router. You can also simply renew their DHCP lease (ipconfig /release followed by ipconfig /renew under Windows). Otherwise, you have to wait until the lease expires and gets renewed by your clients.

 

[CyberMew]

That's what DNSFilter was designed for, yes.

That's great because I need my devices to use the WAN DNS 8.8.8.8 except for a tv box that i have (which does not have a dns setting function, i.e. i think it pulls it from the router)

However I'm not sure if I configured it correctly, since it's still not going through the custom dns server. Would you mind taking a quick look at it?

http://i.imgur.com/L411w90.png?1
http://i.imgur.com/NnvsxHL.png?1

edit: found my issue, lol sorry!

 

[Chrysalis]

Hi again everyone.

Summary of current issues I have here.

First thanks to john for adding the mtu fix to radvd for ipv6 connections.

So here is list.

1 - when making adjustments on the ipv6 page and hitting apply, the router sometimes will become unresponsive, in windows network settings, the router gets listed as unknown and lan traffic to it stops working. Rebooting the router works, and the setting is applied after the reboot. I had this issue even before you started the fork so is an old problem but didnt report it as I was not sure if was down to me not starting from a clean config, this time however I did use a clean config so hence this report, also I dont think all settings cause the issue, it seems to only occur when ipv6 is currently active on the router.
2 - radvd caused an issue on my router when trying to connect to a broken gateway on my isp, so an explanation is this, my isp's new gateways do not work properly with ipv6, its their fault, but the gateway I connect to is random so sometimes I will connect to a broken one, when this happens radvd is hanging whilst trying to start and this hangs the entire /sbin/init process, when using the service command such as to restart dnsmasq, it will fail with a message saying "waiting for radvd to start" and init process is using 99% cpu all the time. As a workaround I disabled radvd on the ipv6 configuration page and will manually start it as when I need it and when I know I am on a good gateway.
3 - The restore config feature seems broken, so I came back to merlin firmware after using shibby for a bit, and installed the same version I last used so matched the version I had when I backed up the config, when I restored the config only a few settings were restored, but about 70-80% were not restored. So I pretty much did everything from scratch (I reset and setup again).
4 - when stopping openvpn client it is not clean. So my openvpn config applies some static routes when started, looking at the log these apply cleanly, however when stopping I see lines like this in the log (this is only a few of the lines)
"Apr 11 16:55:58 rc_service: httpd 322:notify_rc stop_vpnclient1
Apr 11 16:55:58 openvpn[6450]: event_wait : Interrupted system call (code=4)
Apr 11 16:55:58 openvpn[6450]: /usr/sbin/ip route del 149.255.98.15/32
Apr 11 16:55:58 openvpn[6450]: ERROR: Linux route delete command failed: external program exited with error status: 2
Apr 11 16:55:58 openvpn[6450]: /usr/sbin/ip route del 217.156.128.0/17
Apr 11 16:55:58 openvpn[6450]: ERROR: Linux route delete command failed: external program exited with error status: 2
Apr 11 16:55:58 openvpn[6450]: /usr/sbin/ip route del 208.91.156.0/22
Apr 11 16:55:58 openvpn[6450]: ERROR: Linux route delete command failed: external program exited with error status: 2
Apr 11 16:55:58 openvpn[6450]: /usr/sbin/ip route del 199.127.192.0/22
Apr 11 16:55:58 openvpn[6450]: ERROR: Linux route delete command failed: external program exited with error status: 2
Apr 11 16:55:58 openvpn[6450]: /usr/sbin/ip route del 195.27.0.0/16"


Thanks I am now on the latest stable version 374.43_2-10j9527 (Merlin fork)

 

[Trebuin]

I tried to trace back this msg and I think it's a symptom, not a cause. It looks like the router is reacting to a state change action that the repeater has been disconnected and it enters a disconnect/cleanup sequence.

Is the 2.4 GHz band particularly 'busy' in your environment where there could be interference with the AP?

Update: disabling the IPV6 firewall prior to enabling repeater seems to have worked for 2.4GHz. I just switched over to 5GHz to give that a try, will update in about a day or two.

 

[coldwizard]

...
Update-11B2 HighLights:

• New Fork Updates
  • New options (Note: Fork unique options are designated by '**' in the gui)
    • Added gui option to allow local subnet forwarding (special case)

...

Could you explain this "local subnet forwarding" option maybe with an example? How is it different than a static route on the LAN - Route tab?

 

[john9527]

Could you explain this "local subnet forwarding" option maybe with an example? How is it different than a static route on the LAN - Route tab?

See this post....
http://www.snbforums.com/threads/fork-update-for-374-43-available.18914/page-41#post-157359

From the reading I did, it's basically a quirk in INVALID state processing that can sometimes hit packets it shouldn't.