02-04-2015 02:01 AM - edited 03-01-2019 07:47 AM
Hi,
I installed a couple of Nexus 9396PX (NX-OS standalone) configured in vPC.
Unfortunately, as requested by the customer they where not the root bridge of STP (MST has been used) because the core switches are a couple of HP they are connected to in square shape through portchannel (then the Nexus ones collect other switches or UCS below as usual through vPC member ports.
That means that one of HP switches is the root bridge and one link from the Nexus has the port in STP blocking.
Now, I've a couple of questions:
1) We know that normally the VPC switches should be root bridge of STP and that STP makes the peerlink always UP and in forwarding; in my scenario, how could be possible that it happens however? ...I mean, I could think for instance to low the STP cost of the port from the Nexus towards the HP no root bridge and at that point STP could force the peerlink to get in blocking state, right? Or there is a hidden mechanism that override the normal STP computation?
2) All the VPC port member are 10Gbps interfaces towards the access switches/UCS; in accord with cisco doc "MST uses the long path-cost calculation method" and that means:
Bandwidth | Short Path Cost Method Port Cost | Long Path Cost Method Port Cost |
---|---|---|
10 Mbps | 100 | 2,000,000 |
100 Mbps | 19 | 200,000 |
1-Gigabit Ethernet | 4 | 20,000 |
10-Gigabit Ethernet | 2 | 2,000 |
Now, what i see is:
VPC_Nexus_1# sh spanning-tree <-- is the NExus directly connected to HP core root bridge
MST0000
Spanning tree enabled protocol mstp
Root ID Priority 4096
Address 001b.3f55.8e00 <-- is the HP switch
Cost 0
Port 4195 (port-channel100)
Hello Time 2 sec Max Age 20 sec Forward Delay 15 sec
Bridge ID Priority 16384 (priority 16384 sys-id-ext 0)
Address 58f3.9ca3.6825
Hello Time 2 sec Max Age 20 sec Forward Delay 15 sec
Interface Role Sts Cost Prio.Nbr Type
---------------- ---- --- --------- -------- --------------------------------
Po1 Desg FWD 200 128.4096 (vPC) P2p <-- VPC towards the access switch
Po2 Desg FWD 200 128.4097 (vPC) P2p <-- VPC towards the access switch
Po10 Desg FWD 1000 128.4105 (vPC peer-link) Network P2p <-- peer link
Po100 Root FWD 200 128.4195 (vPC) P2p <-- VPC towards the HP_1
Po200 Desg FWD 200 128.4295 (vPC) P2p <-- VPC towards the HP_2
The only STP cost that is correct is the one concerning the peer link made of 2 X 10G; i don't understand where come from the "200" value for the vpc channels!
Thanks for your feedback,
Mario
02-05-2015 06:21 PM
Hi Mario,
Could you please share your network diagram and configuration on these switches, so that I can help you further.
Best Reagrds
Sachi Garg
02-09-2015 03:58 AM
Hi Sachi,
thanks for your support, in attachment you find the configuration of both N9K and the scheme.
The dotted links towards HP switches are because not physically connected but just pre-configured for when they will be in place there.
I'm looking forward to reading your email.
Best regards,
Mario
01-16-2017 04:04 PM
I know this is a really old post, but I've configured a VPC link with a couple of Nexus switches that are also not setup as root and am seeing a 2 Gbps VPC uplink to the root result in a cost of 200. Correct me if I am wrong, but the long method for a 2Gbps link should be 10,000 right? What's going on here?
Only 3 switches in the environment. Cisco 2960 is root, and two Nexus switches with a 20Gbps peer link between them. Spanning tree reports that link with a cost of 1000. The uplink from the two Nexus switches is a VPC 2Gbps link to the 2960.
04-04-2017 01:31 AM
Hello,
"vPC links have a predefined cost of 200. This cost is hard-coded and does not depend on the number of links in the PortChannel."
Have a good day,
Vincent
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: