Ethernet Line Protocol Up Down

Unanswered Question
Apr 13th, 2010

Hi Guys,

All throughout my network I'm seeing the line protocol on our switches go up and down lik a yoyo. Numerous ports on numerous switches have expeirenced the issues and I am struggling to ascertain why.

I cannot produce the issue on demmand, it is intermittent to a particular port but is an active issue across the network.

Some interfaces have errors, of various types, others have none. So I'm assuming the errors come from past issues since the counters have never been reset.

Does any one have any suggestions as to how I best troubleshoot this issue futher?

Thanks in advance

Rgds

Scott

I have this problem too.
0 votes
  • 1
  • 2
  • 3
  • 4
  • 5
Overall Rating: 0 (0 ratings)
Loading.
Leo Laohoo Tue, 04/13/2010 - 18:29

Does any one have any suggestions as to how I best troubleshoot this issue futher?

You can start by telling us what is your switch and IOS you are running.  You may need to post the config of the port having problem.  
Scott Cannon Tue, 04/13/2010 - 19:32

Hi leolaohoo,

The problem is widesprea, completely throughout the network consisting of 2950's, 2960, 3750 and 6500's. Weve got COS and IOS with varying versions. (8 and up for COS and 12.2 and up for IOS)

Rgds

Scott

lamav Tue, 04/13/2010 - 19:08

Scott, this behavior is typical when you have speed and duplex mismatches. Take one pair of ethernet connections and hard-code both sides, just to be safe, and see if that clears the problem.

HTH

Victor

Scott Cannon Tue, 04/13/2010 - 19:34

Hi Victor,

Thanks for the prompt response. I have already hard coded settings on the switches but I will organise for a couple of choice switches to have the endpoints hard coded aswell.

I have an open ear to any other suggestions people might have.

Thanks

Rgds

Scott

Leo Laohoo Tue, 04/13/2010 - 20:31

Except for 2950/2940 (and earlier models), switches support "test cable-diagnostic tdr interface " command to test the integrity of your copper cable.  Try this as well.  NOTE:  Depending on your IOS, this can be disruptive.

Scott Cannon Tue, 04/13/2010 - 22:52

Thanks leolaohoo. I'll do soem investigations and let you know how I go.

Rgds

Scott

ohassairi Tue, 04/13/2010 - 22:19

1-are you using port security feature?

2-did you notice any spanning tree log messages?

what log messages are you getting in console when port goes down?

Scott Cannon Tue, 04/13/2010 - 23:00

Hi ohassairi, and thnks for contributing.

We have port security enabled in small parts of teh organisation but not on all switches experiencing the issue.

There are no spanning tree log messages and no recent topology changes (ie. for months on the handful of sample switches I reviewed)

The syslog events I'm seing are simialr to this on each switch (with differing ports on each switch):

192313: *Jul 11 06:50:45.336: %LINEPROTO-5-UPDOWN: Line protocol on Interface FastEthernet0/2, changed state to up
192314: *Jul 11 06:51:18.429: %LINEPROTO-5-UPDOWN: Line protocol on Interface FastEthernet0/2, changed state to down
192315: *Jul 11 06:51:20.433: %LINEPROTO-5-UPDOWN: Line protocol on Interface FastEthernet0/2, changed state to up
192316: *Jul 11 06:53:04.763: %LINEPROTO-5-UPDOWN: Line protocol on Interface FastEthernet0/2, changed state to down
192317: *Jul 11 06:53:06.767: %LINEPROTO-5-UPDOWN: Line protocol on Interface FastEthernet0/2, changed state to up
192318: *Jul 11 06:53:39.726: %LINEPROTO-5-UPDOWN: Line protocol on Interface FastEthernet0/2, changed state to down
192319: *Jul 11 06:53:41.731: %LINEPROTO-5-UPDOWN: Line protocol on Interface FastEthernet0/2, changed state to up
192320: *Jul 11 06:55:26.027: %LINEPROTO-5-UPDOWN: Line protocol on Interface FastEthernet0/2, changed state to down
192321: *Jul 11 06:55:28.032: %LINEPROTO-5-UPDOWN: Line protocol on Interface FastEthernet0/2, changed state to up
192322: *Jul 11 06:56:17.130: %LINEPROTO-5-UPDOWN: Line protocol on Interface FastEthernet0/2, changed state to down
192323: *Jul 11 06:56:19.152: %LINEPROTO-5-UPDOWN: Line protocol on Interface FastEthernet0/2, changed state to up
192324: *Jul 11 06:58:03.431: %LINEPROTO-5-UPDOWN: Line protocol on Interface FastEthernet0/2, changed state to down

Ignore the time stamp, that particular switch is configured to use NTP. at the time of typing this message the time on the switch Ipuled that log from is: *07:27:21.851 UTC Sun Jul 11 1993

So as you can see the logs are very recent.

Rgds

Scott

bhisham84 Wed, 04/14/2010 - 06:14

Hi,

What is connected to Fas0/2?

Plz check the speed and duplex as well?

plz share sh run & sh int status of this interface.

Thanks/bhisham

Scott Cannon Wed, 04/14/2010 - 15:16

Hi Bhisam,

Thanks for your input. There is a PC on the end of this port, it is set to auto negotiate, which I have since changed to 100 full. I will continue to monitor.

Note: This issue is widespread throughout our network. I wouldnt think we would have to hard code the speed and duplex settings on all devices in the network. Also note, the switch is not advising of any duplex mismatch issues and I've never witnessed a switch pot in half duplex mode (mind you all switch ports are hard coded to full)

Is anyone aware of any known duplex negotiation issues with either NICs and/or switches?

Thanks again

Rgds

Scott

shailesh.h Wed, 04/14/2010 - 07:37

Hi Scott,

First thing that you have to isolote whether it is network problem or problem with the system connected to that port. Few steps which i could recommend is

1. Get one working system and swap the port and observe the result

2. Compare the configuration of working port and non-working port

I suggest draw a rough diagram.. then try to identify topology issue if any (may be L-2 or L-3)..

Share the result to progress...

With regards,

Shailesh Hardas

Scott Cannon Wed, 04/14/2010 - 15:21

Hi Shaliesh,

As an intermitent issue I'm not sure which end hosts are suffering and which arent.

The interfaces dont reset as such so I cant tell from the interface stats. I guess I'm going to have to pick a few choice switches and send syslog messages to a server since the siwtches memory isnt large enough to hold any more than an hours worth of log messages.

I'll come back with some detail once I've collected it.

Open to suggestions in the mean time.

Thanks

Rgds

Scott

charlesdf22 Wed, 04/14/2010 - 19:00

We ended up receiving a bad batch of patch cables that were showing this in the logs.  I have seen this error with cat5 patch cables pushing the limits, crappy punch jobs at the jack and distance.

Doing the test cable diagnostics mentioned earlier (only works on gig ports) will definitely help you figure out if you have bad cables.  Keep in mind that is in intrusive.  If you don't have a gig switch in that closet, a GLC-T on the sfp port should work as well.

Actions

This Discussion