I have an email in with support regarding this. We are seeing very odd OSPF (?) related problems between two of our routers. One is a 1100AH running 5.18, and one is a 450G running 5.14. They are connected via UBNT M5's in WDS mode.

* The two routers have an established OSPF broadcast session between them.
* Several times per day, R1 reports that it's ethernet link goes physically down.
* The OSPF neighbor goes down
* The ethernet link comes back up, but only at 10M/full
* The ethernet link goes back down
* The ethernet link comes back up, this time at 100M/full
* OSPF reports "state change from Exchange to 2-Way" 4-6 times
* OSPF then reports "wrong peer state, state=ExStart"

The only way to get the OSPF session to re-establish is to physically reboot the radio on one end. EVERYTHING on these towers has been replaced - routers, all cabling, UBNT radios, etc. This continues to happen on this one link.

This may not be OSPF related, but at this point, I'm looking for any advice.

Thanks.

you are a victim of the port-flapping issue.

have you tried to upgrade the routerboot?
/sys routerboard upgrade

this fixed the port flapping for me in one case.

No, I haven't, but I will!

Thanks,

Josh

Support told me this was normal for auto-negotiation.. I think not. Anyhow, I upgraded the firmware this morning, so we'll see how it works.

Josh

This RB ran all weekend with no problems. Thanks again!