pmp450 connected, but doesn't pass traffic

Over the last few months, we've had a few SMs that are connected to the AP, but no traffic can be passed either way.  The unit won't respond to pings, but it is still registered to the AP.  The two ways I've found to fix this are to force the SM to disconnect from the AP several times, until it does respond, or to powercycle it.

Almost all of our 450 5ghz cambium is using 13.2.1.

Is there a watchdog feature available in this firmare that will help with this?  If it's there, I haven't been able to find it.

Have you consider upgrading to the 15.0.1 or the latest 15.0.2?

The following link will take you to the firmware download/Release notes,perhaps you could find the answer in the latest firmware versions

https://support.cambiumnetworks.com/files/pmp450/


@acfreema wrote:

Over the last few months, we've had a few SMs that are connected to the AP, but no traffic can be passed either way.  The unit won't respond to pings, but it is still registered to the AP.  The two ways I've found to fix this are to force the SM to disconnect from the AP several times, until it does respond, or to powercycle it.

Almost all of our 450 5ghz cambium is using 13.2.1.

Is there a watchdog feature available in this firmare that will help with this?  If it's there, I haven't been able to find it.


This is rare, but has been happening for a very long time and on every release back to 12.x. Something with the affected SM's session gets stuck. I haven't been able to figure out what triggers it.

The easiest way to fix it is to drop that SM's session at the AP. Go to Tools > Sessions. Select the SM. Select the local option. And then click the drop selected session button. The SM will see the lost session after 30 seconds or so, then rescan and register again. No reboots required.

That said, I haven't seen this on 15.0.2 yet, but it has only been running across our network for a week. So we'll have to wait and see.

According to other coworkers' notes, this happened on several older versions too, but I didn't notice it.  I had a feeling it would be something that is chronic across versions, but I couldn't find any other information about this happening (at least, it wasn't described in a way that seemed similar).

Should I conclude that there isn't a handy watchdog feature for forcing a reboot if some particular ICMP communication doesn't succeed?