CBS250-48PP-4G / CBS250-24PP-4G PoE Problem

[AntonisD]

Hello,

We have one CBS250-24PP-4G and two CBS250-48PP-4G switches. All of them are updated to the latest firmware available (3.1.1.7).

On all 3 of them PoE fails in a strange manner. On ports that have nothing connected on them and even on ports that we have set to disabled, PoE stops working after some time, without any message generated in logs (at debug level).

By that we mean that if a PoE device is connected on the port and a port is re-enabled (we have tried several different PoE devices), nothing happens, not even in logs. If a non-PoE device is connected, the port works normally.

The Operational Status of the port in the webui appears at that moment appears as “Fault”.  By issuing the command Show Power Inline the output appears to be random. Port status is always off but the reason can be anything (Short condition, Power budget exceeded, etc). Also the counters on all ports are random, some of them are 0 some show some value.

In the logs, the only related message that appears sometimes is

Informational    %HAL_config_poe-I-OVLPort: PD connected to Port gi** requires more power than it should. Please disconnect it 

NOT on ports that Delivering Power but only on empty ports no matter if they are disabled, or have PoE disabled or not, and that message alone, does not correlate immediately with PoE not working on the port. The port may as well work fine until it randomly fails, without any message, sometime later.

Ports that actually deliver power on connected devices seem unaffected, until their cable is disconnected and connected again, where sometimes they might go to the off state as well, without generating any related message in logs. (Other than the normal link Up/Down messages)

The only way to make disabled ports to work again is to reboot the switch, or to change Power Mode from Port Limit to Class Limit and vice versa where actually what happens is that the system turns the power OFF and ON for all PoE ports.

The port cannot be restored even if PoE is turned off and on by using power inline never and then power inline auto or by using shut and no shut on the port to disable and enable it completely.

All 3 switches are installed in different environments, on different power distribution panels and UPS. We have tried factory reset several times, and tested all 3 switches outside of our production environment without any useful conclusions.

I found a similar issue in topic https://community.cisco.com/t5/small-business-switches/cbs250-8p-e-2g-poe-odd-behaviour-sometimes-fw-3-1-1-7-unifi-ac/td-p/4511172 where the latest firmware seems to have fixed it. In the firmware release notes actually is stated as fixed "After awhile, the PoE will stop the power supply on some interfaces." but we still have the problem.

Other than the obvious question, which would be a way to fix the problem completely, does anyone know a way to force a port

OFF and back ON without having to reboot the switch or to toggle all ports?

[AntonisD · Accepted solution]

Hello,

after several factory resets due to testing, the switch was left at dynamic DHCP since we have a static mapping with its base mac address at our DHCP server and didn't bother to change it to static on the switch side. Then because I was getting the following warning.

Informational %BOOTP_DHCP_CL-I-DHCPRENEWED: The device has been renewed the configuration on interface Vlan 1 , IP 192.168.86.212, mask 255.255.255.0, DHCP server 192.168.86.254

I changed it back to static IP address, and in combination with the latest firmware (3.1.1.7 for the CBS250 models) the problem was resolved. For testing purposes I changed it back to dynamic DHCP and the problem was repeatable.

[Simplicity]

Hi Antonis, we are experiencing the same issue with CBD250-24PP-4G; our firmware version is 3.1.1.7 and we've also statically assigned an address to the switch. I'm wondering, did you experience this issue again since you last posted and if so did you find a resolution? We have a couple of these needing to go out to a customer environment in the next week and we're nervous about this issue seeing as they heavily rely upon PoE. We've only been able to reproduce the issue by unplugging a PD and plugging back in again, and we're trusting that the customer won't be doing that often and crossing our fingers that this won't become a big issue.

[AntonisD]

Hello, 

We haven't seen the issue since I last posted and actually on one switch, we have a POE WiFi access point that it is used as "portable". It's being disconnected and connected on different ports frequently (and through long cable runs and on wall mounted LAN ports), as it is used in classrooms as a dedicated access point for certain applications.

It looks like your case is a bit different in the sense that POE does not fail randomly on unused ports but only upon PD disconnection. Does this generate any messages in logs? And with what procedure you can restore POE operation? (For example rebooting, resetting the specific port, or by changing power mode as I mentioned in my first post). Since you mentioned that they rely heavily on POE, how many ports are utilized is also important since this is a partial POE switch and I am thinking that this could be an overload issue?