Asus RT-AC87R/U - FACTORY F/W Bug Thread

[lilstone87]

Does anybody have this problem of the 5Ghz LED light constantly going off and back on again (in less than a split second)?

I kept getting this with all the factory firmwares that I tried and even with Merlin's build.

Is my set faulty or is it a bug? Thanks.

This is normal, as there are two different drivers between 2.4GHz band and the 5GHz band. You can disable the LED's if it bothers you.

 

[mlai]

On all the firmwares so far, when setting up ipv6 6in4 tunnelling, the ac87u ignores the manual dns settings, and broadcast its ipv6 address as dns server via router advertisement and dhcpv6. Worse, it does not respond to dns queries on its ipv6 address.

So the local clients all have problems resolving ipv6 addresses. any solution ? I had to manually input the ipv6 dns setting at every clients that i have now to make ipv6 work

 

[raphaelogy]

Some bugs i found on 2678 firmware:

1) Even though the WPS is disabled on GUI, pressing the WPS button on the router enables WPS even when it was previously disabled.

2) Pressing the reset button on GUI under WPS tab resets all wireless settings for 2.4 and 5ghz to default values.

3) using inSSIDer, 5ghz is broadcasted as WPA2 + WPS even though WPS is already disabled on GUI, while 2.4ghz reflects WPA2 normally.

not sure if you guys experience the same..

 

[cc666]

Some bugs i found on 2678 firmware:

1) Even though the WPS is disabled on GUI, pressing the WPS button on the router enables WPS even when it was previously disabled.

2) Pressing the reset button on GUI under WPS tab resets all wireless settings for 2.4 and 5ghz to default values.

3) using inSSIDer, 5ghz is broadcasted as WPA2 + WPS even though WPS is already disabled on GUI, while 2.4ghz reflects WPA2 normally.

not sure if you guys experience the same..

On mine I do NOT see WPA2 + WPS only WPA2 on the 5ghz band.
Same for the 2.4ghz band.

CC

 

[raphaelogy]

On mine I do NOT see WPA2 + WPS only WPA2 on the 5ghz band.
Same for the 2.4ghz band.

CC

Hallelujah.. what shall i do next? Tried rebooting, reset, reflash firmware but still not reflecting what it should be..

Edit: Ok managed to get the WPS turned off by changing the authentication method for 5Ghz from WPA2-Personal to Open system first, then change it back to WPA2-Personal. This method solved my WPS wrongly activated for 5ghz despite GUI displayed WPS as off for both 2.4 and 5.

 

[mgpeters01]

STP - Thank you.

Have you tried disabling spanning tree protocol (STP)

Their isn't any options to access STP settings, I can find. I do have 3 un-managed switches (none of them daisy chained). The only other thing I have that is unusual is I am running ethernet over cable (MoCA) to get a hard signal to the back of my house.

 

[RMerlin]

Their isn't any options to access STP settings, I can find. I do have 3 un-managed switches (none of them daisy chained). The only other thing I have that is unusual is I am running ethernet over cable (MoCA) to get a hard signal to the back of my house.

The setting is only available on the webui of my firmware. You can however manually disable it over telnet on Asus's stock firmware. Enable telnet, connect to your router using a telnet client such as XShell4 or putty, and run these commands:

nvram set lan_stp=0
nvram set lan1_stp=0
nvram commit
reboot

 

[raphaelogy]

I mentioned before about my 5ghz broadcasting WPS feature though it has been turned off at web ui, now finally i found a way to replicate the problem.

The bug is when the router is rebooted, the problem comes back again. In order to remove WPS, I have to turn it on and off again at web ui. Once inreboot the router, WPS is seen again..

Has anyone encountered this bug?

 

[lilstone87]

Well I notice two thing's I see with this newest firmware still .2678.

1. Under traffic monitor, if you select past 24 hours. There is no tab for the 5GHz band, yet under Real-Time, 5GHz band is listed. I know the two wireless bands have different drivers, but not showing a 5GHz band graph for past 24 hours seems to me as a bug. As only Primary WAN, Wired, and 2.4GHz band, have shown usage for past 24 hours.

2. This is a issue I have notice with my 5GHz wireless bridge for a while now, with both newer, and older firmwares. If I select channel's like 40-44 as my main channel, after a period of time my bridge will no longer stay connected. But using what I would consider the "edge" channel's(36/48) in the lower set of channel's it stays connected alright. Still to me this shouldn't be an issue, and I am 99.9% sure this wasn't an issue with previous router's I used, with this wireless bridge.

 

[DaveMishSr]

Well I notice two thing's I see with this newest firmware still .2678.

1. Under traffic monitor, if you select past 24 hours. There is no tab for the 5GHz band, yet under Real-Time, 5GHz band is listed. I know the two wireless bands have different drivers, but not showing a 5GHz band graph for past 24 hours seems to me as a bug. As only Primary WAN, Wired, and 2.4GHz band, have shown usage for past 24 hours.

2. This is a issue I have notice with my 5GHz wireless bridge for a while now, with both newer, and older firmwares. If I select channel's like 40-44 as my main channel, after a period of time my bridge will no longer stay connected. But using what I would consider the "edge" channel's(36/48) in the lower set of channel's it stays connected alright. Still to me this shouldn't be an issue, and I am 99.9% sure this wasn't an issue with previous router's I used, with this wireless bridge.

Try switching your wireless channel setting to "Auto" instead of using a fixed channel. I have had no issues with my 5 media bridges since abandoning using fixed channels on the 5GHz band. Hope this helps.

Good luck!

 

[scavenger]

With Adaptive QoS turned on fw 2678 on a 4Mb / 0.5Mb connection, my system log is spammed with QoS class quantum errors:

[...]
Sep 6 08:03:17 kernel: HTB: quantum of class 100023 is small. Consider r2q change.
Sep 6 08:03:17 kernel: HTB: quantum of class 110023 is small. Consider r2q change.
Sep 6 08:03:17 kernel: HTB: quantum of class 120023 is small. Consider r2q change.
Sep 6 08:03:17 kernel: HTB: quantum of class 130023 is small. Consider r2q change.
Sep 6 08:03:17 kernel: HTB: quantum of class 140023 is small. Consider r2q change.
Sep 6 08:03:17 kernel: HTB: quantum of class 150023 is small. Consider r2q change.
Sep 6 08:03:18 kernel: HTB: quantum of class 160023 is small. Consider r2q change.
Sep 6 08:03:18 kernel: HTB: quantum of class 170023 is small. Consider r2q change.
Sep 6 08:03:18 kernel: HTB: quantum of class 100023 is small. Consider r2q change.
Sep 6 08:03:18 kernel: HTB: quantum of class 110023 is small. Consider r2q change.
Sep 6 08:03:18 kernel: HTB: quantum of class 120023 is small. Consider r2q change.
Sep 6 08:03:18 kernel: HTB: quantum of class 130023 is small. Consider r2q change.
Sep 6 08:03:18 kernel: HTB: quantum of class 140023 is small. Consider r2q change.
Sep 6 08:03:18 kernel: HTB: quantum of class 150023 is small. Consider r2q change.
Sep 6 08:03:18 kernel: HTB: quantum of class 160023 is small. Consider r2q change.
Sep 6 08:03:18 kernel: HTB: quantum of class 170023 is small. Consider r2q change.
[...]

These appear to be related to the automatic classes associated with Adaptive QoS (c.f. http://mailman.ds9a.nl/pipermail/lartc/2003q1/007508.html). I'm thinking A-QoS is broken for any broadband speed less than a certain amount (1 Mb up?) and simply hasn't been tested for these speeds.

 

[RMerlin]

With Adaptive QoS turned on fw 2678 on a 4Mb / 0.5Mb connection, my system log is spammed with QoS class quantum errors:

[...]
Sep 6 08:03:17 kernel: HTB: quantum of class 100023 is small. Consider r2q change.
Sep 6 08:03:17 kernel: HTB: quantum of class 110023 is small. Consider r2q change.
Sep 6 08:03:17 kernel: HTB: quantum of class 120023 is small. Consider r2q change.
Sep 6 08:03:17 kernel: HTB: quantum of class 130023 is small. Consider r2q change.
Sep 6 08:03:17 kernel: HTB: quantum of class 140023 is small. Consider r2q change.
Sep 6 08:03:17 kernel: HTB: quantum of class 150023 is small. Consider r2q change.
Sep 6 08:03:18 kernel: HTB: quantum of class 160023 is small. Consider r2q change.
Sep 6 08:03:18 kernel: HTB: quantum of class 170023 is small. Consider r2q change.
Sep 6 08:03:18 kernel: HTB: quantum of class 100023 is small. Consider r2q change.
Sep 6 08:03:18 kernel: HTB: quantum of class 110023 is small. Consider r2q change.
Sep 6 08:03:18 kernel: HTB: quantum of class 120023 is small. Consider r2q change.
Sep 6 08:03:18 kernel: HTB: quantum of class 130023 is small. Consider r2q change.
Sep 6 08:03:18 kernel: HTB: quantum of class 140023 is small. Consider r2q change.
Sep 6 08:03:18 kernel: HTB: quantum of class 150023 is small. Consider r2q change.
Sep 6 08:03:18 kernel: HTB: quantum of class 160023 is small. Consider r2q change.
Sep 6 08:03:18 kernel: HTB: quantum of class 170023 is small. Consider r2q change.
[...]

These appear to be related to the automatic classes associated with Adaptive QoS (c.f. http://mailman.ds9a.nl/pipermail/lartc/2003q1/007508.html). I'm thinking A-QoS is broken for any broadband speed less than a certain amount (1 Mb up?) and simply hasn't been tested for these speeds.

Those also appear with traditional QoS, and have been there for years.

 

[bodean]

Those also appear with traditional QoS, and have been there for years.

RMerlin: Do you use the Traditional/Adaptive QOS on this router? Or the Apps Analysis? If so, how have they been running for you? I know a few FW's ago users reported issues.

 

[RMerlin]

RMerlin: Do you use the Traditional/Adaptive QOS on this router? Or the Apps Analysis? If so, how have they been running for you? I know a few FW's ago users reported issues.

I use both Adaptive QoS and Apps Analysis here. There was a VoIP-related issue that Asus/Trendmicro resolved with a signature update about a month ago, and the vlan2 issue that was resolved with 2361/2678.

I haven't actively tested QoS since the beta days. Back then, I tested a torrent running at the same time as a Speedtest.net run. The instant I started the speedtest, torrent speed dropped close to 0, and the speedtest accurately reported my full up/down speeds. Once the speedtest completed, torrent went back to its initial 2-3 MB/s.

I've had one occasion one night where suddenly web browsing inside a VM slowed down a lot until I restarted the router, but it hasn't happened again since then.

 

[mromero]

I use both Adaptive QoS and Apps Analysis here. There was a VoIP-related issue that Asus/Trendmicro resolved with a signature update about a month ago, and the vlan2 issue that was resolved with 2361/2678.

I've had one occasion one night where suddenly web browsing inside a VM slowed down a lot until I restarted the router, but it hasn't happened again since then.

I've the same experience twice with the latest official Asus firmware. Browsing slowed down to a crawl via wireless. Tried a wired connected box and it was slow as molasses.Managed to log in very slowly.

Rebooting fixes it.

 

[ctonylee]

I'm having these same issues and it's driving me crazy. seems to always be around the AP unit. Works fine for a bot then boots me and have to do reboot. Speed is good when working but this is really becoming a very serious issue for me. Also why did ASUS get rid of the wireless bridge mode that AC68 had? This was a really cool feature and it didn't split my bandwidth in half. I hope they bring this back as that's the setup I prefer.

 

[roldogg]

Dead loop/QoS issue

I edited my reply, the QoS related issue is the dead loop on vlan2 error, which is different.

I'm getting the same messages in my system log, but I not using QoS nor do I have it or Adaptive QoS enabled:

Sep 12 06:39:20 kernel: Dead loop on virtual device vlan2, fix it urgently!
Sep 12 06:39:20 kernel: Dead loop on virtual device vlan2, fix it urgently!
Sep 12 06:39:31 kernel: Dead loop on virtual device vlan2, fix it urgently!
Sep 12 06:39:31 kernel: Dead loop on virtual device vlan2, fix it urgently!

Also, about once a day, something happens and the router doesn't allow any wireless clients to connect. These are some of the logs right before this happens,

Dec 31 18:00:56 dnsmasq-dhcp[1608]: read /etc/ethers - 24 addresses
Dec 31 18:00:56 dnsmasq[1608]: using nameserver 68.105.28.11#53
Dec 31 18:00:56 dnsmasq[1608]: using nameserver 68.105.29.11#53
Dec 31 18:00:56 dnsmasq[1608]: using nameserver 68.105.28.12#53
Dec 31 18:00:58 rc_service: udhcpc 1031:notify_rc start_upnp
Dec 31 18:00:58 rc_service: waitting "stop_upnp" via udhcpc ...
Dec 31 18:00:59 miniupnpd[823]: shutting down MiniUPnPd
Dec 31 18:00:59 rc_service: udhcpc 1031:notify_rc stop_ntpc
Dec 31 18:00:59 rc_service: waitting "start_upnp" via udhcpc ...
Dec 31 18:01:00 miniupnpd[1685]: HTTP listening on port 35673
Dec 31 18:01:00 miniupnpd[1685]: Listening for NAT-PMP/PCP traffic on port 5351
Dec 31 18:01:00 WAN Connection: WAN was restored.
Dec 31 18:01:00 rc_service: wanduck 601:notify_rc start_apps_update
Dec 31 18:01:00 rc_service: udhcpc 1031:notify_rc start_ntpc
Dec 31 18:01:00 rc_service: waitting "stop_ntpc" via udhcpc ...
Dec 31 18:01:02 kernel: Dead loop on virtual device vlan2, fix it urgently!
Dec 31 18:01:02 kernel: Dead loop on virtual device vlan2, fix it urgently!

and what I find strange is the date of the log. Once I power cycle the router, it starts working again, for about 18-20 hrs but after the router comes back online after I power cycle it, the date is correct,

 

[RMerlin]

I'm getting the same messages in my system log, but I not using QoS nor do I have it or Adaptive QoS enabled:

Upgrade to 376_2678 which resolves this issue. The engine isn't used only for QoS.

 

[Crazy_RT]

I have found a bug in the 5 GHZ WPS toggle or setting.

When you switch WPS off for both frequencies, 5GHZ WPS only remains off until the next reboot or power cycle.

Upon the restart, the GUI will show WPS off, but it is in fact ON on 5GHZ band.

I have confirmed it by using it on m laptop to connect.

The solution is to toggle on and off the WPS until next reboot when it happens again.

This is my 4th RT-ac87. My first one held the setting for some odd reason. But after 3 with the same problem, i assume its a bug.

I am unsure if this gets eventually forwarded to asus, but if anyone knows how to tell them about bugs let me know. Its a little unsettling to have WPS on when its just some numbers, aside from all the buzz about its insecurity. Thats one of the first things I disable always.

J
J

 

[raphaelogy]

I have found a bug in the 5 GHZ WPS toggle or setting.

When you switch WPS off for both frequencies, 5GHZ WPS only remains off until the next reboot or power cycle.

Upon the restart, the GUI will show WPS off, but it is in fact ON on 5GHZ band.

I have confirmed it by using it on m laptop to connect.

The solution is to toggle on and off the WPS until next reboot when it happens again.

This is my 4th RT-ac87. My first one held the setting for some odd reason. But after 3 with the same problem, i assume its a bug.

I am unsure if this gets eventually forwarded to asus, but if anyone knows how to tell them about bugs let me know. Its a little unsettling to have WPS on when its just some numbers, aside from all the buzz about its insecurity. Thats one of the first things I disable always.

J
J

Hi bro. I have already mentioned your problem which I encountered way before you. You might want to look at #63 of this thread.

Actually, doing a reset by pressing the hard button does not solve the issue as well, perhaps this is something that is software related and is pending for resolution..