Nebula Reports AP ( NWA50AX ) Offline, Netconf connection is disconnected

Alan_Twigg1951

Netconf connection is disconnected, followed by Netconf connection is connected.

Nebula marks the AP as offline, however clients are still connected and do not lose the connection. This situation lasts for a couple of minutes then all is well again.

I have 2x NwA50AX configured in a Mesh scenario, 1 AP is connected to my router ( BT Smart Hub ) via ethernet ( 1GB ), the other AP is a Mesh Repeater and connects to the master AP via 5Ghz wi-fi, currently this connection shows as uplink signal level -59

rate: 864/648

I see this error on both of my AP's, daily and both AP's have static ip addresses with Google DNS, 8.8.8.8. I have replaced all my ethernet cables between the AP and router.

When I get this offline error? I am able to telnet to d.nebula.zyxel.com on port 4335 from a linux device connected to the same ethernet switch as my main AP.

Both AP's firmware are , V7.10(ABYW.4)

Any help appreciated

Alan

Zyxel_Tina

Hi @Alan_Twigg1951,

Welcome to the Zyxel Community!

From the logs of your device, we found that the AP repeatedly disconnects from and reconnects to NCC due to RPC timeout.

Based on past experience, this typically happens as follows: NCC communicates with the AP at regular intervals. If there is packet loss on the uplink network, the AP may not receive the RPC polling in time, and will assume the connection session with NCC has timed out — at which point it disconnects and re-establishes the connection via Call Home. In other words, this is mainly related to the overall network environment, meaning the internet path to the Nebula server may be unstable.

Therefore, we'd suggest checking on the ISP and router side for any anomalies, for example confirming whether session limits are being applied, or capturing packets to check for packet loss.

Alan_Twigg1951

So I contact my ISP (BT in the UK ) and say "You have an unstable network" please fixit?

I really dont think they will respond to this, it would be far simpler to relax your " rpc polling window" or at least provide a "switch" within nebula to cater for "slow" networks, job done.