WS-12-250DC Turning off all POE but not rebooting

[stevebarnes35]

It has been really cold here the last 3 days. -15 this morning. I have one WS-12-250DC switch at the top of a water tower that is randomly turning off the POE to all devices on the tower. I have 11 ports used 10 are giving out 24VDC POE.
In the log at the identical same time all ports that have POE say:

Jan 1 09:47:54 Port: link state changed to 'down' on port 6
Jan 1 09:47:54 STP: set port 6 to discarding
Jan 1 09:47:54 Port: link state changed to 'down' on port 7
Jan 1 09:47:54 STP: set port 7 to discarding
Jan 1 09:47:54 Port: link state changed to 'down' on port 8
Jan 1 09:47:54 STP: set port 8 to discarding

But the switch never reboots and the devices connected to the SFP and Non-POE ports stay active.

3 seconds later all the devices start to come back up.

Jan 1 09:47:57 Port: link state changed to 'up' (100M-F) on port 7
Jan 1 09:47:57 STP: set port 7 to discarding
Jan 1 09:47:57 Port: link state changed to 'up' (100M-F) on port 6
Jan 1 09:47:57 STP: set port 6 to discarding
Jan 1 09:47:57 Port: link state changed to 'up' (100M-F) on port 8


Voltage in is 24 VDC in the cold it dips sometimes to 21VDC but no where close to the Hibernate voltage of 11VDC.





Current Time
1/2/2018 11:35:54 AM


Uptime
131 days 00:16:60


CPU Usage
100%590%


Memory Usage
128MB52.50MB


MAC Address
ec:13:b2:81:d4:f4


IP Address
10.100.50.3


Firmware Version
1.4.7


Board Rev
F


Power Supply Firmware Version
53


Power Supply Board Rev
B

[Julian]

Just to clarify, can you make dead nuts sure that it's all PoE ports? I have an suspicion that it may just be 24v PoE ports, which matters.

[sirhc]

Please upgrade firmware to v1.4.9
There was IMPORTANT fixes in v1.4.8 and in v1.4.9

v1.4.8 had an IMPORTANT FIX for all SMART DC models
v1.4.9 has an IMPORTANT FIX for UBNT discovery protocol.

Please always upgrade to the latest firmware before opening a ticket/post requesting help so we do not chase a problem that may have been fixed with a firmware upgrade.

Not saying the firmware upgrade is this issue but why not start the diagnostics processes from a known point with all known issues solved.

Please post up screen grabs of the following TABs:
Status
Ports
Power
Device/Status

[stevebarnes35]

I will upgrade firmware and see if that resolves. I upgraded the power to 48Vdc and It has not had this issue. (i dont think) since.

[mike@surfglobal.net]

We saw the same thing and posted on one of the forums here recently. It only happens on a loaded netonix. For us it seemed to be units with AC Prism radios. The cold pushes the current draw up and trips something inside the switch that quickly drops all 24V POE voltage for a brief instant. We have subsequently put all AC Prism back into their own POE and are pulling all our Netonix switches as time permits at main sites. Between this issue and the problem with AirFibers blowing easily on these switches they just aren't reliable enough for us. They are a great concept, but when an issue happens you lose all your AP's and links connected to it. The standard response on here is always its a grounding issue yet other switches and UBNT POE's work flawlessly. This problem with it dropping all voltage is a reason why using these in a high availability setting is bad. All POE voltages should be isolated so one port doesn't affect another.

[sirhc]

We saw the same thing and posted on one of the forums here recently. It only happens on a loaded netonix. For us it seemed to be units with AC Prism radios. The cold pushes the current draw up and trips something inside the switch that quickly drops all 24V POE voltage for a brief instant. We have subsequently put all AC Prism back into their own POE and are pulling all our Netonix switches as time permits at main sites. Between this issue and the problem with AirFibers blowing easily on these switches they just aren't reliable enough for us. They are a great concept, but when an issue happens you lose all your AP's and links connected to it. The standard response on here is always its a grounding issue yet other switches and UBNT POE's work flawlessly. This problem with it dropping all voltage is a reason why using these in a high availability setting is bad. All POE voltages should be isolated so one port doesn't affect another.

I am sorry to tell you the above is not a correct assumption but you are entitled to your opinion.

I run a WISP and power over 30+ airFIBER radios and see NONE of these issues. If your blowing airFIEBER radios you have poor grounding and probably are not bonding your ac service grounds to your tower ground rods. Since UBNT bonds DC negative to the rail (earth ground) if you do not have good grounding you blow out Ethernet port in the radio and or the switch.

We repair a lot of airFIBER radios as explained in this post: viewtopic.php?f=21&t=2827#p19482

The damage we see to AF and MIMOSA radios and not all repaired units are from people who use our switches are from poor grounding.

There is no way our switches can blow radios.

Your 24V story on airMAX AC radios also does not make sense with the overwhelming number of people who use our switches around the world including me at my WISP. There are over 50,000 Netonix switches in service around the world.

We also supply our switches to Verizon Wireless who uses them to power their mico cells in stadiums and hotels and they love them.

But as I said your entitled to your opinion and sorry your experience was bad with us but I am not sure your issues were related us or our switches but that is my opinion as well.