Switch ports dropping randomly

[rayamitchell]

%LINEPROTO-5-UPDOWN: Line protocol on Interface FastEthernet0/10, changed state to up

%LINEPROTO-5-UPDOWN: Line protocol on Interface FastEthernet0/10, changed state to down:

I'm getting a ton of these messages on random ports during random times. Clients are losing connectivity and sometimes regaining connectivity immediately, while other times the port is bouncing numerous times before connectivity is restored.

Background: (8) 3500XL switches, (1) 2600 series router - providing routing and DHCP

Please give me input as to why this could be happening? It can't be a bad cable or nic, as it is happening in multiple locations with multiple computers/cables.

[Amit Singh]

Ray, Have you recently changed some configuration on these switches related to syslog? I have seen some older

cases where the syslog confgurations changes has done this kind of behaviour. If yes, could you please try disabling the syslog related configs on these switches and see if that changes the behaviour. These switches are EOL/EOS now, i wonder if there is any IOS available to fix this.

Cheers,

-amit singh

[rayamitchell]

I have recently enabled syslog on all switches, but the reason I enabled and setup a syslog was because this issue was occuring so I'm not sure the syslog itself is the cause.

Also, can I update firmware even though they are far beyond their years?

[Amit Singh]

Are they running the same IOS? Ok, first of all disable the syslog settings on these switch and see if stablize the network and this behaviour.

List the IOS version from these switch and lets check if there is an IOS available which is later than the one these are running.

Cheers,

-amit singh

[rayamitchell]

12.0(5)WC3b is what they are running on. One of them is on 12.0(5)WC17.

[rayamitchell]

Also, I have disabled logging on all switches and router.

[Amit Singh]

Ray,

c3500xl-c3h2s-mz.120-5.WC17.bin is the last IOS software published for the 3500XL switches. You already have that running on one of the switch. Try upgrading the same IOS on all the switches and see if it makes the difference.

HTH,

-amit singh

[Leo Laohoo]

I'm getting a ton of these messages on random ports during random times. Clients are losing connectivity and sometimes regaining connectivity immediately, while other times the port is bouncing numerous times before connectivity is restored.

If the clients log out do the interface negotiate to 10/Half?

What is the NIC brand/manufacturers of these PC?

If the answer to your first question is YES and the 2nd question is Intel then I am leaning towards a bug between the NIC and Cisco.  The NIC, a 10/100/1000BaseTx, driver has an issue negotiating to "jitter" with the 3500XL only.  This issue doesn't affect the 2900XL.  This issue has been identified as early as 2008.