In a previous post I worked through enabling trunking and NIC teaming for my VMWare ESX server connected to my Cisco Switch. When I was complete I left behind a problem and the comments here, and here suggested I enabled portfast. Today I attempted to fix the problem and enable portfast.

Switch access

Last week I needed to move all my ports from vlan1 to vlan4. When I did this it removed my access to the management ip of the switches. Thanks to the trusty blue serial cables I am still able to access these switches and configure them. So I added an ip address to interface vlan4 and on two of the switches (2950) it put the interface vlan1 in shutdown. On the third switch (2960G) It allows interfaces on both vlan1 and vlan4.

But try as I might I could not get consistent response on these new vlan management ip addresses. Oh and here is the catch, I am (slowly) migrating from a 192.x ip addressing scheme to a 10.x ip addressing scheme. So I took this opportunity to migrate these switches – at least their management ip addresses – to the new scheme. In hindsight, had I not done this I wouldn’t have had a problem.

Turns out I forgot about ip default-gateway. Evidently I had too many beers Wednesday afternoon to think through routing when changing ip subnets. The simple and quick fix was this:

switch#conf t
switch(config)#ip default-gateway 10.0.1.1
switxh(config)#exit

Viola, fixed.

Portfast

Admittedly, I know nothing about portfast so I did my research and discovered it is simple to enable portfast, yet everything screams errors and problems. So I have it enabled, and it is working. I’ll know more when I get my second ESX server deployed and HA in the mix.

switch#conf t
switch(config)#int range gi0/14 - 15 , po1
switch(config-if)#spanning-tree portfast trunk
switch(config-if)#exit
switch(config)#exit

Questions for the crowd out there, Should I enable portfast on both my ports and my port-channel? Also, I enabled portfast trunk. Why did I have to specify that (I know the ports are trunked), does portfast work differently in a trunked environment or is it the concern of network flooding that they add in an extra step for us.