I have read somewhere that the command "vpc peer-switch" only starts "working" when the Nexus pair is the root of the vpc VLAN. Is this correct ? If the Nexus pair is not the root for some VLANs, what happens if i configure this command ? Will it break something ?
For example in a migration scenario from VSS to vpc Pair. The two environments are interconnected by back-to-back L2 portchannel. However, in the beginning of the migration, the VSS is still the "root" of the VLANs.
The Nexus pair will behave as "one" switch, as per default, however only primary will process BPDUs. The received BPDUs on the secundary will be forwarded to the primary for processing.
What happens when i configure "vpc peer-switch" in this environment ? i guess nothing changes as long as the Nexus is not made root, it continues to work in the primary/secundary roles ? correct ? Even if i make priorities the same, it still continues to work in primary/secondary roles ? (ie the secundary will not generate BPDUs by itself). ONLY when this VLAN becomes Root, will the secundary start generating its "own" (same) BPDUs ? and the switch can make this differentiation automatically PER VLAN ? is that correct ? It also means that from the moment it becomes "root" the system-ID (of this vlan) will change from the hardware system-ID to the logical system-ID ? and only for this vlan then ?
What is the general recommendation here ? is it recommended to make the VLAN immediatly root on the Nexus BEFORE migrating switches from VSS to the Nexus pair, or is it recommended to keep it as non-root, then move all switches to the Nexus pair, and in the end make all vlans root ? What has the least amount of downtime ?