Cisco Catalyst 4000 series Hardware Troubleshoot page 15

It is always recommended to keep the management or sc0 interface of the switch in a separate VLAN
from the user data. Otherwise, any broadcast or multicast storm can flood the inband channel to the
Network Management Processor (NMP), which needs to be free to handle the protocols just
mentioned.
If you have not been able to track down any reason for performance issues on the switch in the
previous steps mentioned, capture the output of these commands, as well as the other commands in
the previous steps, and open a case with the Cisco Technical Support:
♦
show nvramenv 1 (hidden)
♦
show interposition 1 (hidden)
These are hidden commands, which means they cannot be parsed with a "?" and you cannot Tab to
complete. Type the command out in its entirety. It is possible that this output is not useful in the
resolution your case. These commands are undocumented, and therefore the TAC engineer is not
required to explain the output to the customer.
7. Although fairly rare, memory leaks do occur and can cause what seem naturally to be poor
performance and other symptoms. If you have not been able to track down any reason for
performance issues on the switch in the previous steps mentioned, capture the output of the show
mbuf total (hidden) command, as well as the other commands in the previous steps, and open a case
with the Cisco Technical Support.
There are two things to consider when you look at the output of this command in order to help
determine if you have a memory leak issue:
♦
Look at the output and if the free mbufs or clusters values decrease but never increase, this
can indicate a possible memory leak.
♦
Look at the output, and if the lowest free memory has ever approached zero or was at zero,
this indicates the switch either runs low on or has ran out of memory.
Both of these issues indicate a memory issue that obviously affects the protocols/processes that
require this memory.
Cat4000−c> (enable) show mbuf total
mbufs
free mbufs
lowest free mbufs
These are hidden commands, which means they cannot be parsed with a "?" and you cannot Tab to
complete. Type the command out in its entirety. It is possible that this output is not useful in the
resolution your case. These commands are undocumented, and therefore the TAC engineer is not
required to explain the output to the customer.
Getting system error messages
As mentioned in the Introduction of this document, Cisco has a suite of online diagnostic tools to help you
determine hw/sw compatibility, interpret output, and decode errors. One of these tools is the Error Message
Decoder. This tool requires a login and can be used to decipher the output of system messages you are
concerned about.
1. System messages have timestamps by default, which can help in isolating a timeframe for your
problem. by Issue the show time command in order to make sure your system clock is set correctly.
Also, verify that your connecting devices are set so that the logs match.
2. Capture the output of any system messages with the show logging buffer 1023 command. Many
system messages are informational in nature while others can indicate a problem. Refer to these
documents for more information:
9280 clusters
9256 clfree
9235 lowest clfree
3660
3659
3638