I just upgraded all of our switches on campus to Version 15.0(2)SG4 after about a month of testing. On two switches so far, we are seeing that clients can not connect, and the switch isnt detecting a link. I dont see anything out of the ordinary in int status, port-security, or errors on the interface. Plugging in a different computer does nothing. Only thing that works, is a shut, no shut of the interface. After that, its connected. anyone else run into this issue?
On the switches that are having issues, every other port seems to be working fine, or at least we haven't heard of issues. All of the switches have the same sups, WS-X45-SUP6-E. There does not seem to be any operating differences between switches that are working, and ones that we have had issues with.
Is it possible the rommon is too old on the ones that don't work correctly ? Something to check if you haven't already.
You must upgrade to at leaset ROMMON Release 12.2(44r)SG5 to run Cisco IOS Release 15.0(2)SG on the Supervisor Engine 6-E and Supervisor Engine 6L-E. 12.2(44r)SG9 is recommended.
Im having the same issue.. We upgraded our closets to 4510s Sup7 . 17 of the 4510s are not having any client issues.. 4 of them are.. They are all same rev/hw type on all the swithes.. What makes the problem worse is the ports dont go into shut down and it makes it a little more work to find the issued ports and do shut / no shut to get things working again..
Looks like this is an IOS bug.
|line cards stop responding to interface changes|
One or more line cards stop responding to interface changes and ports remain
down/down (notconnect) when connected to hosts. Interfaces that are already in
the link-up state are not affected.Conditions:
The problem only affects WS-X45-SUP7-E supervisors running 03.01.00.SG,
03.01.01.SG, 03.02.00.SG or 03.02.01.SG.
The line cards that can be affected are in the WS-X46xx series and the WS-X47xx
series. Earlier line cards are not affected.Workaround:
Upgrade to 3.2.5SG (3.2.2SG and higher has the fix).Further information
Avoid triggering the issue by eliminating link flap.
If module < n> is already in the problem state:
- reload the linecard module by doing 'hw-module module < n> reset'.
- if you have dual supervisors, perform a switchover
I do have an update. We tried Version 15.1(1)SG1 and it seems to have the same issues. We are still on ROM
12.2(44r)SG10 and Sup 6-E.
The problem arises when a computer reboots and the link between the nic and the switch drops. Once the computer boots back up, the port never re-connects. You can disable/enable the nic, swap computers, and its still down. The connection is only restored when you do a shut/no shut on the port itself.
Its pretty annoying that I can really only re-create the issue in our production enviroment, and its happened with the last three versions of code. The last tac case just suggested to stay on a old version of code, and TAC didnt seem worried that there was a bug. We seem to be safe on 15.0(2)