PPPoE secondwan sometimes takes several minutes to recover after fiber reconnect

Router
1

Hello GL.iNet team,

I am testing PPPoE on the secondwan interface and noticed an interesting recovery behavior that may be worth investigating for a future firmware update.

After disconnecting and reconnecting the fiber connection, secondwan sometimes takes several minutes to re-establish the PPPoE session.

During the failed attempts, the log repeatedly shows:

pppd: Timeout waiting for PADO packets
pppd: Unable to complete PPPoE Discovery
netifd: Interface 'secondwan' is now down
netifd: Interface 'secondwan' is setting up now

The router automatically retries the PPPoE connection.

Eventually, PPPoE Discovery succeeds. The successful sequence includes:

pppd: PPP session is established
pppd: Connected to the PPPoE peer via interface eth1.2
pppd: PAP authentication succeeded
netifd: Network device 'pppoe-secondwan' link is up
netifd: Interface 'secondwan' is now up

After physically rearranging the fiber and reconnecting it again, recovery took approximately 30 seconds.

The connection has now remained stable for approximately 30 minutes with no drops.

This is interesting because the problem is not consistently reproducible. Once PPPoE successfully establishes the session, the connection appears to be stable.

Could the engineering team investigate the PPPoE discovery/retry behavior?

In particular:

  1. Why can the router repeatedly wait for PADO packets for several minutes before successfully discovering the PPPoE server?
  2. Could the PPPoE discovery/retry process be improved after a physical WAN interruption?
  3. Could the Ethernet/PPPoE state be reset or reinitialized more aggressively when the physical link comes back?
  4. Is there anything in the current firmware that could explain the long sequence of PADO timeouts?

I am not claiming that the fiber or ISP is definitely the cause. The main observation is that PPPoE eventually works and remains stable, but recovery can sometimes take several minutes.

I am providing this report so the behavior can be investigated and, if appropriate, improved in a future firmware update.

I can provide the complete system log if needed.

Thank you.

4

Hi,

Thank you for providing the detailed report.

This may happen if the previous PPPoE session has not yet been released by the upstream server.

When the fiber connection is interrupted, the previous PPPoE session may remain on the upstream server because it was not terminated normally. During this period, the upstream server may not immediately respond to a new PPPoE discovery request.
The router will continue retrying automatically, and a new session could normally be established after the previous session times out and is released by the upstream server.

Thank you again for the detailed report and feedback.