Cisco Support Community
cancel
Showing results for 
Search instead for 
Did you mean: 
cancel

Catalyst switch running CatOS system software continuously resets, or crashes.

Core issue

A Switching Bus Timeout (SWB) occurs when there is no switching bus activity for 1.6 seconds. There is a hardware timer on every supervisor card which monitors switching bus signals from the time of installation. If it does not see change at least once in 1.6 s, and generates a Non-Maskable Interrupt to the NMP on the supervisory Module, that resets the catalyst.

Resolution

If the switch has reset or crashed due to a hardware or software issue, it is important to capture the output of certain show commands as quickly as possible.

  • Capture the output of these commands:      

    1. show log       
    2. show version       
    3. show test 0       
    4. show logging buffer 1023
         

    The show log command output indicates these problems that may be related to a crash:

       
    • It keeps track of the last 10 system resets with timestamps, showing when the reboot occurred. The following is a snapshot of the reboot history output:     

      Reboot History:   Jan 23 2002 11:14:16 0, Jan 22 2002 14:57:21 0

                         DEC 24 2001 13:56:38 0, DEC 24 2001 13:52:30 0

                         DEC 11 2001 12:31:59 0, DEC 07 2001 13:26:48 0

                         DEC 07 2001 10:42:19 0, DEC 07 2001 10:36:16 0

                         Nov 28 2001 11:03:10 0, Oct 26 2001 16:04:26 0

           

      The reboot history indicates only that the switch was reset. It could have been reset manually by the user or caused by a crash. However, the most recent manual reset of the switch is recorded further down in the output, as shown in the following example:

           

      Last software reset by user: Jan 23 2002 11:14:16 0

           

      Notice that the timestamp of the last manual reset, 1/23/2002,11:13:13, matches the most recent entry in the reboot history.

             
    • It shows if there have been any exceptions. Exceptions are CPU dumps that occur immediately after a crash, as shown in this example:     

           

      MCP Exceptions/Hang:    0

           

      In the above case, there were no exceptions recorded. If there were an exception, it would include a timestamp that could be matched with the reboot history. It would also include a hex dump or stack which could be decoded by a Technical Assistance Center (TAC) engineer to determine whether this is a software forced exception or due to hardware.

             
     

    The show version command provides software version information to use for a bug search. For example, if you identify an exception in the show log command output, refer to the Bug Toolkit to search for bugs on the Catalyst 4000 and the exception. The show version command gives you a quick snapshot of how long the switch has been up, as shown in this example:

     

    Uptime is 28 days, 11 hours, 42 minutes

     

    The show test 0 command output indicates an F status on the supervisor or module if any of the diagnostics failed. An improperly seated module can cause the switch to crash. If the supervisor or module shows failed status, perform the steps in the System component LEDs are orange/red or supervisor not online section of Hardware Troubleshooting for Catalyst 4000/4912G/2980G/2948G Series Switches.

     

    The show logging buffer 1023 command displays all system messages, including possible error messages that may relate to the crash. Refer to the Getting system error messages section of this document for suggestions.

       
  • Issue the show commands and follow the procedures in the preceding steps first. If these steps fail, capture the show tech-support command output. 

    Often, the output from all these commands is not necessary to resolve a specific problem, so TAC engineers may not ask for it. However, it is beneficial to have this output should other show commands or steps fail to resolve the problem.

     

    These commands are supported by the Output Interpreter tool for Catalyst OS (Cat OS) and can be used to assist in solving hardware issues, as shown:

               
  • Should all the previous steps fail to diagnose the problem, keep all the data that has been captured to provide to the TAC, and open a case with the Cisco Technical Assistance Center (TAC).

For additional information, refer to Common CatOS Error Messages on Catalyst 6500/6000 Series Switches

967
Views
0
Helpful
0
Comments