BE85 ethernet backhaul dropping
Set main parent/gateway unit up and three satellite/child units. First setup was a disaster and had to be factory reset and start from scratch.
Second setup resolved most issues, except for same experience with dropping ethernet backhaul within a few hours on both child units that were plugged into ethernet (third child unit is not wired to ethernet and remains a wireless only unit).
After that much more successful second setup, ethernet backhaul on the two units will show wired/connected in the app for a few hours, then both drop and revert to wireless only. So far I have learned that a system reboot will not address the issue. Yesterday, I went to the two affected child units and pulled their power cord and removed/replaced the ethernet cables. Ethernet backhaul then came back on, yet again dropped after only a few hours.
Please help resolve this issue, so I do not have to return these 4 BE85s
- Copy Link
- Subscribe
- Bookmark
- Report Inappropriate Content
Sorry to hear about the behavior, hopefully I can help provide some insight into what may be causing it.
How are your nodes wired together? Do they pass their traffic through a switch? What Model switch?
When passing through a switch, it is important that the switch supports specific features regarding mesh backhauls. It is also possible, depending on the features of the switch that a network loop is being created when the nodes try to switch to a wired backahul - which then blocks the port at the switch.
If the satellite nodes are wired to the network through a switch, try wiring the nodes directly to each other and seeing if the behavior follows. This will help determine where in the line the communication is breaking.
The Decos will recheck their backhauls in a few situations when a new "active" ethernet connection is found (can trigger with swapping ports) or when a node is rebooted. This is also when the Decos will automatically optimize their wireless connections. If possible, you should also consider checking your ethernet cables for a fault. A fault could introduce a behavior like this, so you may try wiring the nodes directly to each other with a new cable to see if the nodes disconnect.
- Copy Link
- Report Inappropriate Content
Thanks for the feedback. It makes more sense it is a switch issue and not a wire issue. Wire is all Cat 6 and new, plus it happens no matter where its plugged in.
It does pass through a switch. Current network map is cable modem ---> BE85 parent node (modem plugged into 10Gbps outlet 1) ---> switch (araknis AN-210-SW-R-24-POE) plugged into 10Gbps outlet 2 ---> BE85 child nodes
If it's thought to be a switch issue, how do I fix this? Can that switch be configured to work or should I try another certain brand/model (open for recommendations, but I need at least 20 or so ports)?
Additionally, this afternoon when I was continuing to troubleshoot, when I reboot one of the wired nodes to try to get it ethernet backhaul back online, it would knock the other wired node's ethernet backhaul offline (yesterday, as noted in the OP, I was able to get both of their ethernet backhauls online after rebooting the two nodes individually - just for them to vanish 6 or so hours later).
Also brings up another puzzling thing to me - if it is an issue (switch or otherwise) why is it it fine/perfect for usually hours at a time, then it decides to no longer work? Also, this was not an issue with another brand's newer model mesh system (went with TPlLink BE85 to get more speed and, I thought, reliability), so what is making BE85 incompatable with the switch, if that is indeed the issue?
- Copy Link
- Report Inappropriate Content
To troubleshoot, you are going to want to first verify that the switch is the culprit by connecting one of the satellite decos directly to a port on the main deco node.
If the culprit is found to be the switch, I would then turn your attention toward the loopback detection and spanning tree features of your switch. When the backhaul method switches between wired and wireless, a short network loop is technically created(~5s). Any loopback detection features of your switch may be taking effect and blocking the port.
Disabling the loopback detection features for the ports should help, especially in re-establishing their backhauls. This feature is dependent on the switch's features but is likely present.
When the backhaul method changes or needs to reconfigure itself due to a change in environment, IP address, wiring, or any number of factors - a short loop may be seen by the switch. If they are happening at the same time, you may try and find a setting that is causing the nodes to want to reconfigure or refresh their wired backhauls.
- Copy Link
- Report Inappropriate Content
I will contact araknis to try to see how to resolve tomorrow - way over my head and the switch management app/website only lets you configure very basic settings, not STP. Not that I would know where to start anyways. The manual has a number of pages on STP, but nothing on loopback detection specifically.
If we cannot get the araknis switch to work with wired backhaul, does tp-link have any 24-port switches that "play nice" with wired backhaul?
- Copy Link
- Report Inappropriate Content
Information
Helpful: 0
Views: 227
Replies: 4
Voters 0
No one has voted for it yet.