Cisco Redundancy

Hi !

I planned to install redundant Cisco equipments on my Lan. I would like you to give me your opinion about my configuration. Thank you in advance

I have 2 backbone switches on my LAN. CATALYST 4510R.

For active redundancy, i configured first the vtp mode.

4510_prod and 4510_backup are vtp servers in the same domain. the 4510_prod has a bigger revision version. The synchronization works fine.

For virtual ip addresses, i configured HSRP

For example:

4510_PROD:

interface Vlan50 description VLAN TEST ip address 10.33.150.252 255.255.255.0 no ip redirects no ip proxy-arp standby 50 ip 10.33.150.1 standby 50 priority 100 standby 50 preempt

4510_BACKUP:

interface Vlan50 description VLAN TEST ip address 10.33.150.253 255.255.255.0 no ip redirects no ip proxy-arp standby 50 ip 10.33.150.1 standby 50 priority 10

For Spanning-Tree, i think the better way is to install MST mode on all my Cisco switches.

I defined 4510_prod as root and my 4510_backup as secondary root (for backup)

For 4510_prod:

Spanning-tree mst configuration Name TEST Revision 1 Instance 1 vlan 1-999 Spanning-tree mst 1 priority 4096

Question: Does i Need to configure the same thing on all my other switches on my LAN? By default, PVST is active.

I cannot delete instance 0. It is normal?

Thank you very much for your help

Greetings

PILU

Be careful, you seem to be mixing routing redundancy with switching redundancy. Both are beneficial, but each works differently. Make sure you keep them separate in your thinking. MST only works within the domain of a single subnet.

Be careful with VTP. Some people have reported problems with the wrong configuration being propagated when replacing failed hardware. Make sure you understand what the protocol is and how it works.

Standard HSRP config, no problems here. Don't forget that you need a default gateway on every VLAN. Also make sure you're MST configuration is solid at layer two and consistent with how you are approaching layer 3.

Remember to use "portfast" on end-user ports and fast failover on trunks. 801.1q can take some time to detect a failure and negotiate a new spanning tree. Negotiation can be particularly troublesome under heavy load conditions, when fast recovery is most important. Tricks like fast failover only improve availability if you understand how spanning tree works, configure root and link costs appropriately, and maintain the configuration as switches are added/removed and topologies change.

Per VLAN spanning trees allow you to load share trunks and reduce the impact of failures (only the VLANs whose spanning tree has that trunk active will go down and renegotiate a new spanning tree). Per VLAN spanning trees are not (yet) standard, so be careful, particularly in a mixed vendor environment.

All your switches need to implement a consistent MST configuration. Make sure that not only is the default spanning tree correct, but also that any single failures will cause negotiation of an acceptable replacement spanning tree. Assuming default values will give acceptable results is a dangerous assumption, particularly if mixing vendors, or even product lines within a single vendor.

Good luck and have fun!