Netonix Forums

Hello.

I am not touching the switch and i keep getting my rocket M5 radios and other equipment going off and on.
This is the log.
Any reason????? Firmware 1.4.2

That is an unexpected link change alert. The switch is telling you the port speed changed for an unknown reason, beyond that it is up to you to figure out why.

Possible causes:
1) Bad cable and or ends
2) Bad radio ethernet port
3) Radio is rebooting - check up time
4) Grounding issue where the tower is not bonded to the service ground and is injecting errors causing the link speed to change and will eventually damage the port and or device or both.

Here are some good posts on grounding:
viewtopic.php?f=30&t=1816
viewtopic.php?f=30&t=188
viewtopic.php?f=30&t=1429
viewtopic.php?f=17&t=1786&start=30#p13447

I doubt it is the switch port but this is easily tested by putting the cable in another port and if the problem follows the cable to the next port then its not the port.

Well I have manged to find out when it happens but i do not know why.
If i leave everything then there is no problems log stays clear no link state change evey thing is up.
I checked all my cables with a network tester they all show the same db, no shorts and a pass on all.
Even though i get some message about bad cable or end when everything passes on the network tester.

This is when the problem happens.
AirControl2 / Discover / IP-Range Scan / Scan 10.1.1.0/24 / then most of the ports that are powering the rocket m5s changed to red.
If i just dicovery broadcast everything is fine.

So i dont know if scanning an IP range is making the APs reboot or scanning an ip range is making the switch ports reboot???

nightshadow36 wrote:I checked all my cables with a network tester they all show the same db, no shorts and a pass on all.

Even though i get some message about bad cable or end when everything passes on the network tester.

Our cable test reports bad cable or end because the Rocket shorts the un-used pairs 3 and 4 (on 10/100 Mbps Ethernet Port) to carry DC power. The way UBNT did this was not the best/standard way but a lot of vendors did this to save money on manufacturing. If you un-plugged the Rocket we would report all pairs OPEN and the length and report cable test PASSED. The Rocket since it is shorting Pair 3 and Pair 4 is why we report FAILED. What they do by pulling the power directly from the pairs without using a center tap Ethernet transformer is not industry standard but as I said is a common practice.

nightshadow36 wrote:Well I have manged to find out when it happens but i do not know why.

When AirControl2 / Discover / IP-Range Scan / Scan 10.1.1.0/24 / then most of the ports that are powering the rocket m5s changed to red.

If I do just a discovery broadcast everything is fine.

So I don't know if scanning an IP range is making the APs reboot or scanning an IP range is making the switch ports reboot???

That is an interesting question, could you turn POE OFF on a couple ports and use the POE bricks and then do the same scan. If the ports with the POE brick still go red then you know the POE is not being turned OFF.
I am not sure how the scan could affect the switch in anyway but I am always open to verification.

Issue went away when i turned off UBNT discovery

nightshadow36 wrote:Issue went away when i turned off UBNT discovery

Try v1.4.3rc5 and see if you can leave UBNT Discovery on then.

Thank you but it works without discovery on and i do not need it.
No need to test now, thanks again

nightshadow36 wrote:Thank you but it works without discovery on and i do not need it.
No need to test now, thanks again

Well it would be nice if you could find out if v1.4.3rc5 fixes your problem so we can know if the problem is solved or need to look deeper.

Maybe do a test over the weekend or evening after prime time.

Could help others out.