cancel
Showing results for 
Search instead for 
Did you mean: 
cancel
345
Views
0
Helpful
8
Replies
Highlighted
Enthusiast

RPVST+ Configuration

Hello,

 

I am working on a new switch upgrade for our enterprise and wanted to confirm the RPVST+ design. See attache for illustration:

 

    - SW-A is the primary root bridge and assigned priority 4096

    - SW-B is the secondary root bridge and assigned priority 8192

    - Both SW-A and SW-B and interconnected via L2 802.q (Trunk Port)

 

   - SW-C is attached to SW-A and assigned root bridge priority 12288

   - SW-D is attached to SW-B and assigned root bridge priority 12288

   - There is no redundancy created between the switches. 

 

Are you seeing any issue with this RPVST+ design? 

 

Thanks in advance,

~zK

1 ACCEPTED SOLUTION

Accepted Solutions
Highlighted

Thanks for the feedback.

 

The core switches are C9500s and the access switches are C9300s.

 

I agree with your recommendations; however, the client didn't want the two core switches to be configured using StackWise Virtual Stacking and wanted the two switches to be interlinked via a trunk (4x10Gbps) so when they have to perform any type of maintenance on one switch they other switch won't be impacted.

 

The end-devices (tenant routers; production ESXs, Load-balancers, ASA, and other production servers) will be uplinked to the access-switches. Each end-device will have duel-home connections to each access-switch (C9300, that's SW-C and SW-D on the diagram.) That's why the client is fine without having the other links from the access-switches to the core-switches ( redundant links). 

 

You've provided me with enough information to complete my design. Much appreciated!!!

 

Best, ~zK

View solution in original post

8 REPLIES 8
Highlighted
VIP Mentor

High level i do not see any issue and also you do not have any network loops here.. so stratinght forward.

 

until you have any other network connected and we missing here ?

 

BB
*** Rate All Helpful Responses ***
Highlighted

Thanks for your input.

 

Yes, there will be downstream devices connected to SW-C & SW-D. Those devices are end-point devices (ESXs, servers, LBs, ASAs, and routers). Also, I just discovered that there is a single  C3750G switch that will be dual-homed to both SW-C & SW-D. I updated the design and its attached.

 

With that said, do you see any issues with this design?

 

Best, ~zK

Highlighted
VIP Mentor

Hello

not at all - would suggest you add an additional link to either core from swC/D for resilience.



kind regards
Paul

Please rate and mark posts accordingly if you have found any of the information provided useful.
It will hopefully assist others with similar issues in the future
Highlighted

Thanks, Paul!

 

I just update the design and post above. Any thoughts?

 

Best, ~zK

Highlighted

Suggest all access switches to connect to core with dual connection so you have high availability.

 

attached rough diagram.

 

BB
*** Rate All Helpful Responses ***
Highlighted

Wouldn't lose two ports with this design? Meaning, wouldn't two ports be put in a Block state? See attached.

 

 

Thanks, ~zK

Highlighted

The spanning-tree block the alternative path, but advantage with this design is you have resilience link and 100%uptime, with alternative path.

 

Old design have high availability, but if the Core switch go down, you have network outage, 

 

The latest design is better. not sure what switch models is this, if it 6509 you can do VSS / Cat 9500 you can Stackwise virtual / NExus you can to vPC,  in all case you have both the links up and running. 

 

Normal spanning-tree this happends any way of blocking the alternative path. that is reason we need to play balacing the vlans on both the switch example .

 

1,3,5 - odd vlan SW1 as root bridge

2,4,6 even vlan Sw2 as root bridge.

 

BB
*** Rate All Helpful Responses ***
Highlighted

Thanks for the feedback.

 

The core switches are C9500s and the access switches are C9300s.

 

I agree with your recommendations; however, the client didn't want the two core switches to be configured using StackWise Virtual Stacking and wanted the two switches to be interlinked via a trunk (4x10Gbps) so when they have to perform any type of maintenance on one switch they other switch won't be impacted.

 

The end-devices (tenant routers; production ESXs, Load-balancers, ASA, and other production servers) will be uplinked to the access-switches. Each end-device will have duel-home connections to each access-switch (C9300, that's SW-C and SW-D on the diagram.) That's why the client is fine without having the other links from the access-switches to the core-switches ( redundant links). 

 

You've provided me with enough information to complete my design. Much appreciated!!!

 

Best, ~zK

View solution in original post

Content for Community-Ad