Radio becomes non responsive

I have a point to point link that had been very stable for months, but has been going down frequently in the last few weeks. We have replaced the original equipment, thinking there was a hardware problem, but it seems to be traffic related. The radios lose connection, and you cannot ping anything through ether1. However the link light stays up on the switch. A reboot always brings it back, sometimes, for a few days, and other times, for just few hours.

I have turned on remote syslogging, with debug enabled, but it does not show the radio was having any kind of problem before it went down. The only clue that anything went wrong is from the far radio log that shows it is deauthenticating due to extensive data loss, but I expect that if the near end has locked up.

Any idea what type of logging I can turn on, beyond info, warning, critical, wireless, and debug?

Maybe someone else has had the same problem and has a fix??

I have seen these problems. Completely puzzling - especially since hardware watchdog is supposed to reboot a routerboard if it becomes unresponsive.

I’ve only changed the channel.

Greetings!

If you have any entries in /ip route that involve the wireless card, you might want to try setting the /interface wireless setting “disable-running-check=yes”. If there is something interfering with the wireless signal and the connection is lost, the wireless card will show “not running” and the route info associated with that card will show “Invalid”. The docs imply the route may not come back when cards reconnect. I have not tried this yet. You might want to create this situation and see if that is what is happening to your setup. I would like to know the results if you try it.

EDIT: Also this (quote from online docs) may apply to a bridge using WDS:

When the link between WDS devices, using wds-mode=dynamic, goes down, the dynamic WDS interfaces disappear and if there are any IP addresses set on this interface, their ‘interface’ setting will change to (unknown). When the link comes up again, the ‘interface’ value will not change - it will remain as (unknown). That’s why it is not recommended to add IP addresses to dynamic WDS interfaces.

If you want to use dynamic WDS in a bridge, set the wds-default-bridge value to desired bridge interface name. When the link will go down and then it comes up, the dynamic WDS interface will be put in the specified bridge automatically.

I ended up replacing the wireless cards, since we had already replaced the routerboard. It has not gone down for several weeks now, so not sure what the difference is, but glad it is working reliably now.