I have been facing problem with Catalyst 3548XL & 2950T switches as they report wrong port numbers ( id's ) when polled by an Linux server using snmpwalk command.
There are 14 x 3548XL switches connected on Gigabit dot1Q trunk to 3 x 6 port GBIC module in the core 4006 core chassis. Ports 1 to 12 in each switch is configured for default VLAN, and the management ip ( interface VLAN 1 ) is also in the same default vlan.
When a pc is connecetd to port 1 (physical) of the 3548XL switch, it gets an ip address from the DHCP server ( Linux ) present in the same default VLAN connected to the core switch. When SNMPWALK command is issued in the Linux server targeting the switch, wrong port number is reported in the output.
The command used in Linux server is as follows
snmpwalk <ip add of switch> public 17.4.3.1.2
where,
public = community string configured in the switch
17.4.3.1.2 = OID used to poll
The output of the command displays port 13 while the physical port connected to the pc is port 1. This misbehaviour is noticed in all the 3548XL switches while the first port is reported as 7 in 2950T switch. Moreover, a strange behaviour was noticed like after every 8 ports, the sequence is skipped by one number like
Switch Port SNMPWALK output
1 13
2 14
3 15
4 16
5 17
6 18
7 19
8 20
9 22 *
10 23
and so on ...
As listed above after 20, 21 is skipped and port ID is reported as 22 for the 9th port.
Just a single command has been included in all the switches for SNMP as below
switch(config)# snmp community public ro
To confirm the operation of the linux server, we had connected the server to a 3Com switch port 12 & a pc to the port 1 of the same switch and snmpwalk command in the linux server reported the pc to be connected to port 1 correctly unlike cisco switches.
Can you pls help me out with the right solution for this problem so that the snmpwalk reports the right port in the linux server as per the physical port number in the switch( like port 1 reported as 1 instead of 13 ).
