cancel
Showing results for 
Search instead for 
Did you mean: 
cancel
682
Views
0
Helpful
5
Replies

What is best method to migrate from existing core infrastructure

ajamerica
Level 1
Level 1

Our customer has decided to migrate off of a existing 3750 stack and to implement a new Collapsed Core design with dual Cisco 4506s.  We were planning to do a one for one swap with the core equipment, but we were informed that we are allowed the bare minimum of downtime.  We decided to merge the 4506s to the network by using a L2 trunk to the 3750s and adding VTP to move the VLANs over.   We are planning to use HSRP on the 3 core switches to keep the default gateway the same by changing the interfaces on all the VLANs to a .254 with a standby address of .1.  What would be the downtime in moving these ips?  Is this the best method to migrate with little downtime?

1 Accepted Solution

Accepted Solutions

Hello Ajamerica,

Does HSRP have any affect to updating the EIGRP tables?

HSRP is not a routing protocol and provides first hop redundancy to end user devices

so it shouldn't affect EIGRP adjacencies

>> we had the HSRP timers set to msecs and not seconds.  Would this cause the 3750 stack process to spike?

This is possible I would suggest to try with default timers

Hope to help

Giuseppe

View solution in original post

5 Replies 5

Giuseppe Larosa
Hall of Fame
Hall of Fame

Hello Ajamerica,

Note on performance:

depending on how the C4506 switches are populated ( supervisors and linecards) this move could even lead to lower performances in comparison with the C3750 stack.

If your C4506 are not E-model with E linecards you could end up with linecards with 48 GE ports and only 6 Gbps of speed towards the backplane!

>> We are planning to use HSRP on the 3 core switches to keep the default gateway the same by changing the interfaces on all the VLANs to a .254 with a standby address of .1.  What would be the downtime in moving these ips?  Is this the best method to migrate with little downtime?

If I understand correctly you are not using HSRP now with the stack, so in each Vlan/IP subnet end user devices have an ARP entry for the MAC address of the SVI on the C3750 stack that is a MAC address of the stack master switch.

if you enable HSRP a gratuitous ARP is sent by active router and this usually updates the ARP table of end user devices.

A gratuitous ARP is also sent when the active router changes.

Warning: some unix boxes may discard the gratuitous ARP so check your environment.

In tests I have done with windows and linux boxes this approach works well.

However, take in account the first note about performance because really the move can be not a good move.

Hope to help

Giuseppe

Thanks so much for your response!  We have installed 4506-E with SUP6E as well.  Is the gratuitous ARP enabled by default on all Cisco switches?  What is the best way to move EIGRP routing tables over to the 4506s?  All the existing routing and vlans reside on the 3750 stack.

thanks so much for your help!

Hello Ajamerica,

4500 E so yes the move makes sense.

>> What is the best way to move EIGRP routing tables over to the 4506s?

enable EIGRP over two Vlans that span between C3750 stack and 4506_1 and another one between C3750 and 4506_2.

make EIGRP passive on all client Vlans where you enable the HSRP groups

>> Is the gratuitous ARP enabled by default on all Cisco switches?

It is part of HSRP implementation

Hope to help

Giuseppe

This is starting to clear up more and more! Does HSRP have any affect to updating the EIGRP tables? The reason I ask is that we tested this the other night and I saw the HSRP tables update over and over again along with the EIGRP tables...

One note I do need to add is that in the initial testing of the migration, we had the HSRP timers set to msecs and not seconds. Would this cause the 3750 stack process to spike?

Thanks!

Hello Ajamerica,

Does HSRP have any affect to updating the EIGRP tables?

HSRP is not a routing protocol and provides first hop redundancy to end user devices

so it shouldn't affect EIGRP adjacencies

>> we had the HSRP timers set to msecs and not seconds.  Would this cause the 3750 stack process to spike?

This is possible I would suggest to try with default timers

Hope to help

Giuseppe

Getting Started

Find answers to your questions by entering keywords or phrases in the Search bar above. New here? Use these resources to familiarize yourself with the community:

Review Cisco Networking products for a $25 gift card