PoE device knocks out my switchport until I reboot the switch

  • 0
  • 1
  • Problem
  • Updated 3 years ago
  • Solved
Hey guys, I've been having this ongoing issue with my x440-48p's.

I had a contractor come in a few weeks ago, who kept trying to connect a low-end PoE switch to one of my switchports - using one of the powered ports on his switch to uplink.

Immediately upon connecting, the port on my 48p would start blinking amber/green. When he moved it to another port, he knocked out that port too. Affected ports kept blinking amber/green until I rebooted the entire switch. Then they were fine.

Recently, another contractor came in while I was on vacation and did the same thing. I read them the riot act of course, but now I'm curious: Am I going to have to reboot my switches again, or is there a way to clear the error status for a port?

Photo of Jack Kelly

Jack Kelly

  • 80 Points 75 badge 2x thumb

Posted 4 years ago

  • 0
  • 1
Photo of Paul Russo

Paul Russo, Alum

  • 9,694 Points 5k badge 2x thumb
Hello Jack

Did you try to disable the port and disable inline power and re-enable.  In past when a port is in a failed state re-enabling them should bring them back.

What are the errors that the port is seeing?  You can do a clear inline-power stats port to clear the errors.

Let me know if this helps or if you are seeing other issues.  What version of code are you using?

I will try and reproduced.

Photo of Paul Russo

Paul Russo, Alum

  • 9,694 Points 5k badge 2x thumb
Great Jack.

I am still going to test.  If you get a log message one of the things we can look at is doing a dynamic UPM profile where if we see the log message automatically run these commands to bring the port up without user intervention.

 Does that make sense?

Photo of Sumit Tokle

Sumit Tokle, Alum

  • 5,738 Points 5k badge 2x thumb
@ Paul is absolutely correct. Let us know which software running on switch?

What is log messages appeared on switch when this issue happened? I could provide you the small script to reset that port automatically when certain alarms would get generated.
Photo of Khawar Shehzad

Khawar Shehzad

  • 70 Points
I am using 15.3.1 patched version. I am having the same problem. All of my ports went off at the same time :(. They are not coming back. I disabled/enabled the ports but that does not help. I restarted the switch. Did not helped. I rebooted with version 12 but the same problem i.e. ports are still off. 
To check whether my servers have the issue or not, i connected my servers to another switch and the lights were working on the other switch. I reconnected to the switch and the problem is still there. 
I have a single lan and i have auto-neg ON. 
Management port is working with easy ssh2 access, no problem there.
My switch is x450a-48p.
I am doing port mirroring on one port and the monitored and the mirrored port are on the same server, as is required.
I am doing an experiment which sends 1 Million queries for a certain service to a lan connected server within the cluster. All queries are the same e.g. same request and same response relationship.
The lights went off when the experiment was running. Dos-protection is off. 
Love is there, as i like extreme switch.
Any help Paul, any help Sumit. 
Photo of Khawar Shehzad

Khawar Shehzad

  • 70 Points
All the diagnostics have passed BUT i am having the following error in the extended diagnostics

Failed to receive of packet in test 8 loopback 2 iteration 0 error code 14
ERROR: Received 0 out of the expected 1 packets on user port 25 on MAC-0
ERROR: TX packet 1 did not return to CPU.

All ports are in a READY state BUT they are no going into the ACTIVE state EVEN though I unconfigured the switch. 

Any help. Anybody?
Photo of Drew C.

Drew C., Community Manager

  • 40,684 Points 20k badge 2x thumb
Hi Khawar,
Since diagnostics have failed, please contact GTAC for further assistance.