Random switch ports are not working - 6509

[Prince K Mathew]

Hello,

I am facing a strange issue with my cisco 6509 core switch. The issue is that all of a sudden some of the ports would stop working. None of these ports are error disabled. Status would show as connected. It would not draw an ip address from DHCP server. I have tried manually assigning an IP address; still doesnt work. After rebooting, it started working for time being. A day later the issue is back and i am left without an answer. This is the 3rd time am coming across it, first instance ports were some random ones. 2nd and 3rd instance, the ports were identical and they are still down. All these ports belong to same module/slot. 

The issue is almost same as below:-

https://supportforums.cisco.com/discussion/11881186/cisco-4506e-random-port-problem-no-ip

We are running IOS version 12.2(33)SXH8. I am seeing constant changes in spanning tree topology originating from random ports. 

Number of topology changes 73 last change occurred 00:01:29 ago from GigabitEthernet9/18

Strange thing is "show log" doesnt display any logs

Monitor logging: level debugging, 0 messages logged, xml disabled, filtering disabled

Have anyone come across to such a scenario. Any helpful suggestion deeply appreciated. 

Thanks,

Prince Mathew

[paul driver]

Hello

Sounds very strange, Any chance you can you reseat the module or relocate into a spare slot?

As for the stp tcns, do your access ports have stp portfast enable and are they all hardcoded or negotiating at full duplex?

 If any access ports are running half duplex  or portfast isnt enabled then this could also cause lots ot tcns.

res

Paul

[Prince K Mathew]

Hi Paul,

Thanks for your response. I would seek a down time for relocating the module and let you know.

My access ports have not confiured with stp-portfast. They have been set to auto negotiation.

Thanks,

Prince Mathew

[KenWynWilliams · Accepted solution]

Hi Paul,

you need to set all non switch ports to spanning tree port fast as this is well known to cause issues with DHCP requests when a port first connects due to it being in blocking mode until spanning tree has done its stuff.

regards

ken

[Prince K Mathew]

Hi Ken,

I shall give it a try during next down time.

Meantime, have some more observations.

one port is running in 10 Mb, Full duplex mode as below

Full-duplex, 10Mb/s, media type is 10/100/1000BaseT

Also, another port which is having 405394 input errors and runts

5 minute input rate 0 bits/sec, 0 packets/sec
  5 minute output rate 0 bits/sec, 0 packets/sec
     458217 packets input, 11100447 bytes, 0 no buffer
     Received 7175 broadcasts (5829 multicasts)
     403202 runts, 0 giants, 0 throttles
     405394 input errors, 404272 CRC, 86105 frame, 1 overrun, 0 ignored
     0 watchdog, 0 multicast, 0 pause input

Will these have an impact?

Thanks,

Prince Mathew

[paul driver · Accepted solution]

Hello

For the stp tcns issue as I stated previoulsy, I would suggest if you dont have portfast enabled on the access ports then apply it - ether gloablly or interface level, and check your interface status of these acess ports for half duplex settings.

spanning-tree  portfast default
or

int x/x
spanning-tree portfast

show interfaces status

Regards your last post, this could indeed be problematic but first clear the interface counters to see if these interface errors are current as they could be historical.

clear counters
clear counters x/x

res

Paul

[Prince K Mathew]

Hello,

Ever since i have enabled port fast, the stp topology change has limited. Also, there is no local changes as i used to get previously. As of now the ports are working without reseating the module. I am not sure if it had something to do with port-fast too.

Thanks,

Prince Mathew

[Leo Laohoo]

I think you have a bad cable here.  

Post the complete output to the following commands: 

1.  sh interface <BLAH>; and 

2.  sh counter interface <BLAH>