Aimesh nodes becomes offline sometimes.

[akb]

Hello All,

Hope I can get some help. I have a ax86u with merlin latest and 3 XD4 mesh nodes.

Every so often i see the log below and one random mesh node is in the offline state. I am usually able to ssh into the node and send the command "reboot" and it recovers. Other times I need to power cycle the mesh node to recover it. This seems like new behavior as I had the same setup for about a year and only changes were the merlin firmware as the new XD4 firmware always breaks my TPlink Kasa devices so I keep them on a firmware which works. Hope I can get some pointers as my mesh nodes actually extend my network to devices which can not reach the main router so it causes outages.

Apr 11 09:25:32 kernel: potentially unexpected fatal signal 11.
Apr 11 09:25:32 kernel: CPU: 1 PID: 8272 Comm: dcd Tainted: P           O    4.1.52 #2
Apr 11 09:25:32 kernel: Hardware name: 94908REF (DT)
Apr 11 09:25:32 kernel: task: ffffffc02c4bb580 ti: ffffffc02c754000 task.ti: ffffffc02c754000
Apr 11 09:25:32 kernel: PC is at 0xf6ce939c
Apr 11 09:25:32 kernel: LR is at 0x1dd58
Apr 11 09:25:32 kernel: pc : [<00000000f6ce939c>] lr : [<000000000001dd58>] pstate: 600f0010
Apr 11 09:25:32 kernel: sp : 00000000ffbe1778
Apr 11 09:25:32 kernel: x12: 00000000000a2050
Apr 11 09:25:32 kernel: x11: 00000000f5fff024 x10: 00000000000a23c4
Apr 11 09:25:32 kernel: x9 : 00000000f5fff810 x8 : 00000000000a287c
Apr 11 09:25:32 kernel: x7 : 00000000f5fff848 x6 : 00000000000a2876
Apr 11 09:25:32 kernel: x5 : 0000000000000000 x4 : 00000000f5fff7f4
Apr 11 09:25:32 kernel: x3 : 0000000000000000 x2 : 00000000ffbe1754
Apr 11 09:25:32 kernel: x1 : 000000000007d652 x0 : 0000000000000000

 

[Tech9]

Does stock Asuswrt on your Main router solve the issue?

 

[akb]

Does stock Asuswrt on your Main router solve the issue?

I am currently testing an older merlin, which worked before.

 

[RMerlin]

The dcd crash you posted is related to the Trend Micro components, and will not have any impact on AiMesh.

 

[akb]

The dcd crash you posted is related to the Trend Micro components, and will not have any impact on AiMesh.

Thanks for replying to this. Interesting as I do not have the Ai security enabled. Would the crash be from it not being enabled, it actually being enabled even if ui says it is off, or just something that is expected to crash on its own.
Aside from that, i guess I have no clue then why the node would go offline and drop all internet connectivity to the clients connected to the AP, but I am able to ssh into the device to recover it sometimes.

 

[RMerlin]

Thanks for replying to this. Interesting as I do not have the Ai security enabled. Would the crash be from it not being enabled, it actually being enabled even if ui says it is off, or just something that is expected to crash on its own.

Adaptive QoS and Traffic Analyzer are other features that use the Trend Micro engine. If you ver used any of these features in the past, then the engine gets loaded and initialized.

If you are positive you are currently not using any of them, you can go to the Privacy tab to retire your consent to the Trend Micro EULA. That will disable loading any of the Trend Micro engine components.

The dcd crashes is a long-standing issue that is already known by Asus and Trend Micro. It can be caused by multiple sources, one of them being if you run something like pixelserv-tls which adds a virtual network interface to your router - that interface can cause the dcd process to crash.

 

[DTM]

I have a theory on the AiMesh dropouts. I was having issues like I see described many places. I fixed it by deleting and resetting my nodes. The node that was crashing was initially set up on an old firmware and the other nodes were set up on the newer firmware. Note the main router and all nodes had later been updated but the troublesome one was INITIALLY set up on an old version. My theory is that it was missing something that made it unstable.

After deleting all nodes, resetting them, and adding them back to the main router, all is well. I had a heck of time getting them added. ASUS needs to work on that! In the beginning adding a node over wifi worked...not any more it seems.
---
How to add a node: Plug an ethernet cable into the WAN port on the node and a LAN port on the main router. Reset the node by holding the reset button for 15 seconds and then let it come all the way back on. From the home screen (Network Map) on the main router, NOT from the AiMesh page, click on the AiMesh icon below the View List button. An AiMesh panel will open on the right. Use the search button and go from there. That is the only way I ever got a new node added.
---
My network is an AC88U main router with 3 AC68P nodes. I hope this helps somebody. If it works or doesn't work for you please let us know.

 

[Tech9]

I had a heck of time getting them added. ASUS needs to work on that! In the beginning adding a node over wifi worked...not any more it seems.

If your main router runs Asuswrt-Merlin, you have to use a cable to add nodes. Wireless node discovery works on stock Asuswrt only.

 

[jata]

I had a similar issue with one of my nodes. It's a RT-AX58U and was working fine. Then after doing a full WPS reset and adding it to my network after a network rebuild, it started to disconnect for no reason (when in wifi backhaul mode).
The issue was that the 5ghz radio stopped working. Even the light on the unit was out. I removed it from the network. did a wps reset re-added it but happened again.

So if you get this issue again, check the 5ghz radio light on the node to see if it's on or off.

To resolve this issue, I started over but flashed the firmware using the firmware restoration tool. Not sure if this low level reset of the firmware helped or maybe just luck. Who knows but worth a try and it only takes a little extra time.

[Wireless Router] How to use Rescue Mode (Firmware Restoration)? | Official Support | ASUS Australia

Another tip is that I always (and only) use stock firmware on my nodes. In my experience, stock on nodes provides a more stable network (and the other advantage is you don't need to update them very often and not at the same time as upgrading your main router)

 

[DTM]

I had a similar issue with one of my nodes. It's a RT-AX58U and was working fine. Then after doing a full WPS reset and adding it to my network after a network rebuild, it started to disconnect for no reason (when in wifi backhaul mode).
The issue was that the 5ghz radio stopped working. Even the light on the unit was out. I removed it from the network. did a wps reset re-added it but happened again.

So if you get this issue again, check the 5ghz radio light on the node to see if it's on or off.

To resolve this issue, I started over but flashed the firmware using the firmware restoration tool. Not sure if this low level reset of the firmware helped or maybe just luck. Who knows but worth a try and it only takes a little extra time.

[Wireless Router] How to use Rescue Mode (Firmware Restoration)? | Official Support | ASUS Australia

Another tip is that I always (and only) use stock firmware on my nodes. In my experience, stock on nodes provides a more stable network (and the other advantage is you don't need to update them very often and not at the same time as upgrading your main router)

In my case and others I have read about the node would disconnect for about a minute and then reconnect. I hope I don't have the problem again but if so I will try to check the 5 GHz light.

 

[DTM]

If your main router runs Asuswrt-Merlin, you have to use a cable to add nodes. Wireless node discovery works on stock Asuswrt only.

Thanks, I was not aware of that. I have always run Merlin on my routers so back when I set up my first AiMesh nodes they were added via wifi with no issues. Needing to use a cable must have come along with one of the firmware revisions.

 

[Tech9]

Thanks, I was not aware of that.

I believe it started doing that some time after switching to 386 code base and AiMesh 2.0 with first reports of something broken coming from AC68U variants owners. At this point wired node discovery workaround is perhaps the only option for most router models used as nodes, if the main router is running Asuswrt-Merlin. Most AiMesh users report better results with stock Asuswrt on nodes. There is no need for Asuswrt-Merlin there anyway.