Dismiss Notice

Welcome To SNBForums

SNBForums is a community for anyone who wants to learn about or discuss the latest in wireless routers, network storage and the ins and outs of building and maintaining a small network.

If you'd like to post a question, simply register and have at it!

While you're at it, please check out SmallNetBuilder for product reviews and our famous Router Charts, Ranker and plenty more!

AiMesh on the T-Mobile TM-AC1900

Discussion in 'ASUSWRT - Official' started by DHume, Oct 20, 2017.

  1. badbob001

    badbob001 Occasional Visitor

    Joined:
    Oct 30, 2012
    Messages:
    16
    Why would an aimesh node with ethernet backhaul go offline? I have three ac68u (one main, two nodes) and they all have ethernet backhauls. On the main unit, I was testing lowering the tx power from max to balance.

    I then checked the wifi analyzer app on my phone to see how that affects the signal strength of my wifi. Then I notice that one of the three duplicate ssid is on a different channel from the others, which is unexpected. I check the admin page and sure enough, one of the nodes has become offline.

    But I can still ping the ip of it through the lan so why would it be considered offline? For each of my nodes, I already have set 'connection priority' to 'ethernet'.

    I had to poweroff and on the offline node to get it connected again.

    The client list shows wifi connections from my two nodes to the main unit, which is also strange since it doesn't need that since it has the ethernet backhaul.
     
  2. badbob001

    badbob001 Occasional Visitor

    Joined:
    Oct 30, 2012
    Messages:
    16
    I found an asus faq that says to use the wan port for the backhaul.
    https://www.asus.com/us/support/FAQ/1035140/
    "Run Ethernet cables from the LAN port of the AiMesh router to the WAN port of AiMesh node."
     
  3. camedia

    camedia New Around Here

    Joined:
    Jan 7, 2018
    Messages:
    7
    I had these problems, but after upgrading to 20308 they are mostly gone.
     
  4. Ronald Schwerer

    Ronald Schwerer Regular Contributor

    Joined:
    Jan 8, 2017
    Messages:
    152
    Of course a wired back-haul goes from a router LAN port to the node's WAN. I thought you and I were talking about the initial set-up of the node. Your link is for after the setup (see step 1).
    What I have not found documented (but know it sometimes works) is the setup step via wired (or it might actually be a wired/wifi hybrid). After failing many times to get a prospective node to successfully finish the setup (using the documented method) , I finally wired it as above and performed node reset. Same location as before - right next to the router. But it finally succeeded all the way. Then I moved the node to the desired location.
    Another way that works is to configure the prospective node as an AP, then log-in and change it to an AiMesh node. Follow the AP webui setup screens, at some point it actually mentions using a wired connection to complete the setup. This was an older firmware version so the verbage might have changed.
     
  5. badbob001

    badbob001 Occasional Visitor

    Joined:
    Oct 30, 2012
    Messages:
    16
    Huh. You're the second person to imply that connecting to the node's wan port is an obvious thing. I was thinking the nodes were using the wifi as a wireless bridge. And since everything in the mesh is on the same subnet, why can't the backhaul be through the lan ports? It's not as if I have a direct cable from the node's wan port to the main unit's lan ports. The wan port is connected to a lan switch that connects to the lan ports on the main unit.

    I also assumed the wan port would be bridged to the lan port as it is in AP mode.

    Yes, I was talking about the initial setup. But given how new and fluid aimesh is, I wouldn't be surprised if this config is also used from the start. Have you tried adding a node via the wan-to-lan port method with WIFI disabled on the node via the external button? If it works, then it would prove that it can use the cable to add nodes.
     
  6. fifty50

    fifty50 Occasional Visitor

    Joined:
    Jan 29, 2018
    Messages:
    14
    thanks, i'll give this a try. default log level is 6, i'll bump to 7.

    second question, system log shows all MAC addresses.. possible to change to device name for an easier read?
     
  7. Ronald Schwerer

    Ronald Schwerer Regular Contributor

    Joined:
    Jan 8, 2017
    Messages:
    152
    I doubt it, most logging is done in low-level routines with just basic input. Converting MACs to friendly names is a much higher level function. It would also add more overhead to the CPU if that lookup was used in more places than the GUI.

    BTW, I wouldn't be surprised if the logging debug level went to higher settings. May even "go to eleven". (Spinal Tap)
     
    Last edited: Feb 8, 2018
  8. fifty50

    fifty50 Occasional Visitor

    Joined:
    Jan 29, 2018
    Messages:
    14
    Figured that may be the case with friendly names.

    So the last few days, my Nest camera that's within 5ft of the node has been going offline for atleast 10min (i don't get alerts until it's been offline for > 10min). usually when I check, the node shows as offline as well. it just happened, and when I connect to the ASUS app, it shows the node is offline, but that the camera is still 'connected' to that node, but the camera is offline since i'm unable to access it. shouldn't it be dropping and connecting to another node (i only have two in AiMesh).

    I'll have to wait until I get home, as I haven't been able to get WAN access to the asus router. My setup is currently FiOS Router (with WiFi radios off), hardwired to ASUS main router which broadcasts my wifi.
     
  9. badbob001

    badbob001 Occasional Visitor

    Joined:
    Oct 30, 2012
    Messages:
    16
    It's likely that even though the node is 'disconnected' from the main mesh, it still accepts client connections. So in essence, it has become of blackhole for those clients.

    It's still a mystery why a node gets disconnected from the mesh. I have a node with an ethernet backhaul that became offline from the mesh when I was adjusting wifi settings (power, channel). I could still ping the ip of the node and trying to go to the its admin page will still redirect me to the page of the main unit.

    Checking the wireless analyzer on my phone, I see that the 'offline' node is on a different wifi channel from the others. When I change the wifi channel of the mesh to match the offline one, it came back into the mesh. So it looks like the nodes are not trying hard enough to get reconnected to the mesh. And perhaps node synching relies too much on wifi; maybe the ethernet backhaul is limited to just client data. I do have the nodes configured to prefer ethernet, but still not convinced if that means that 100% of the mesh data goes over it.
     
  10. fifty50

    fifty50 Occasional Visitor

    Joined:
    Jan 29, 2018
    Messages:
    14
    Ok, so I looked at the logs. the camera went offline around 6:43a. Here are the logs around that time, and went the camera came back.

    f8:32:e4:XX:XX:XX is the node
    18:b4:30:XX:XX:XX is the camera

    thoughts? i also updated to 384_20308 this afternoon too.

    Feb 8 06:07:09 roamast: eth1: add client [40:4e:36:XX:XX:XX] to monitor list
    Feb 8 06:44:32 dnsmasq-dhcp[280]: DHCPREQUEST(br0) 192.168.2.19 40:b4:cd:XX:XX:XX
    Feb 8 06:44:32 dnsmasq-dhcp[280]: DHCPACK(br0) 192.168.2.19 40:b4:cd:XX:XX:XX amazon-1d432XXXX
    Feb 8 06:44:34 roamast: eth2: add client [40:b4:cd:XX:XX:X] to monitor list
    Feb 8 06:50:16 dnsmasq-dhcp[280]: DHCPREQUEST(br0) 192.168.2.19 40:b4:cd:XX:XX:XX
    Feb 8 06:50:16 dnsmasq-dhcp[280]: DHCPACK(br0) 192.168.2.19 40:b4:cd:XX:XX:XX amazon-1d432XXXX
    Feb 8 07:01:29 kernel: br0: port 4(wds1.1) entering forwarding state
    Feb 8 07:01:29 kernel: device wds1.1 left promiscuous mode
    Feb 8 07:01:29 kernel: br0: port 4(wds1.1) entering disabled state
    Feb 8 07:01:30 kernel: device wds1.1 entered promiscuous mode
    Feb 8 07:01:30 kernel: br0: topology change detected, propagating
    Feb 8 07:01:30 kernel: br0: port 4(wds1.1) entering forwarding state
    Feb 8 07:01:30 kernel: br0: port 4(wds1.1) entering forwarding state
    Feb 8 07:01:31 dnsmasq-dhcp[280]: DHCPDISCOVER(br0) 192.168.2.105 f8:32:e4:XX:XX:XX
    Feb 8 07:01:31 dnsmasq-dhcp[280]: DHCPOFFER(br0) 192.168.2.105 f8:32:e4:XX:XX:XX
    Feb 8 07:01:31 dnsmasq-dhcp[280]: DHCPREQUEST(br0) 192.168.2.105 f8:32:e4:XX:XX:XX
    Feb 8 07:01:31 dnsmasq-dhcp[280]: DHCPACK(br0) 192.168.2.105 f8:32:e4:XX:XX:XX
    Feb 8 07:01:51 dnsmasq-dhcp[280]: DHCPDISCOVER(br0) 18:b4:30:XX:XX:XX
    Feb 8 07:01:51 dnsmasq-dhcp[280]: DHCPOFFER(br0) 192.168.2.185 18:b4:30:XX:XX:XX
    Feb 8 07:01:51 dnsmasq-dhcp[280]: DHCPREQUEST(br0) 192.168.2.185 18:b4:30:XX:XX:XX
    Feb 8 07:01:51 dnsmasq-dhcp[280]: DHCPACK(br0) 192.168.2.185 18:b4:30:XX:XX:XX
     
  11. crifever

    crifever New Around Here

    Joined:
    Feb 9, 2018
    Messages:
    4
    I am struggling with this one. I have 2 TM-AC1900 which I have successfully converted to Asus RT-AC68U by following the process documented on this thread. I have the latest Asus firmware installed on both: 3.0.0.4.384_20308 which has the AiMesh functionality. The issue however is when I try to add the node to the AiMesh: it finds the node, I click on adding it to the mesh and it goes thru the syncing process (0% to 100%) in about 2 to 3 minutes and then it gives me this failed message (see attached screenshot)
    Screen Shot 2018-02-08 at 11.50.01 PM.png

    that is clear as mud!!! WTF is going on.... would appreciate some insight on this.....
     
  12. badbob001

    badbob001 Occasional Visitor

    Joined:
    Oct 30, 2012
    Messages:
    16
    I had the same issue and what worked for me was to add an ethernet cable from the node's wan port to a lan port on the master unit. I suspect it'll still work if the lan connection goes through another lan switch on it's way to the wan port.

    I'm curious if the initial syncing can occur over the cable only. You can test this by disabling wifi on the node first with the external wifi on/off button.
     
  13. MysticGold04

    MysticGold04 Occasional Visitor

    Joined:
    Jan 6, 2018
    Messages:
    24
    Location:
    Troubleshooting
    Did you flash the AIMesh CFE on your Main router? That is a requirement if you have converted TM-AC1900's.

    I had no problems linking mine - wirelessly- one upstairs, one downstairs with no issue.
     
  14. crifever

    crifever New Around Here

    Joined:
    Feb 9, 2018
    Messages:
    4
    I will give this a try this evening.
     
  15. crifever

    crifever New Around Here

    Joined:
    Feb 9, 2018
    Messages:
    4
    Well initially I did not use the Aimesh CFE and it wouldnt event show me the node to add. After I put the Aimesh CFE on both - main router and node, it shows me the node to add, goes thru the process of adding the node but then displays that message I had posted earlier.
     
  16. MysticGold04

    MysticGold04 Occasional Visitor

    Joined:
    Jan 6, 2018
    Messages:
    24
    Location:
    Troubleshooting
    Try clearing the NVRAM by either holding the WPS button down and powering on until the power light flashes quickly, or go into the GUI and do a factory reset, let it boot and then leave the node powered up. Try placing the router and the node in close proximity while trying to pair.
     
  17. crifever

    crifever New Around Here

    Joined:
    Feb 9, 2018
    Messages:
    4

    I tried what you recommended with the ethernet cable and wifi enabled, the result is the same. It finds the node, does the 3 minute sync from 0 to 100% and then show the failure message.

    Also attempted to do the initial sync over cable only by disabling the wifi on the node. With the wifi disabled and the node (WAN port) connected to the router thru ethernet cable, the router does not even detect an AImesh node. So you need to have wifi enabled to do the initial sync.


    I am about to give up and use the damn 2nd router as an Access point for my 2nd floor. Cost of the routers $45x2=$90. Energy spent trying to get this cheap mesh network going: f**king priceless. Oh well, you win some you lose some and you curse at those that dont work out the way you intended. Got to love technology!
     
  18. Nebulaz

    Nebulaz New Around Here

    Joined:
    Jan 25, 2017
    Messages:
    5
    I have a question and am a little embarrassed as I have not figured out why on my own. But I have read about 250 posts in this thread and can't figure out what AiMesh is accomplishing.

    If you have to run a cable to the node why or what is different from setting up an access point with roaming assistant on?

    If you don't run a cable, how is it different from a repeater? Do you not lose half of the bandwidth?

    Thank you



    Sent from my LG-H830 using Tapatalk
     
  19. Ronald Schwerer

    Ronald Schwerer Regular Contributor

    Joined:
    Jan 8, 2017
    Messages:
    152
    I'm not sure what my quoted post has to do with your question. The pros/cons of AiMesh have already been discussed ad nauseum in other threads. If you still don't see a benefit, it isn't for you. Don't worry, at this stage you aren't missing much except frustration.
     
  20. Kyle753

    Kyle753 New Around Here

    Joined:
    Jan 5, 2018
    Messages:
    3
    Update from my last post. I flashed both routers to the newest firmware 3.0.0.4.384.20308 and cleared NVRAM on both host and node and the node connected successfully like before. Only difference is now the node no longer drops overnight! Also devices transition between one router to the next when I go relocate my devices (top floor vs basement). Overall I'm very happy with the new firmware, great job Asus!
     
    arthurlien and MysticGold04 like this.

Share This Page