Hello. I am new to networking.
Recently there is a need to add a new vlan to one of the instance in the mst configuration among the switches that run MSTP. And I have searched and read post that there must have some network interruption when configuring the switches. I would like to know in what way the interruption would be minimized when configuring the switches - from root to non-root or from non-root to root and why is it like that.
Thanks very much!
---Do not forget to rate useful post---
I have read this post before, but seems it does not quite answer my question. It uses the VTP server. But in my environment the VTP method is not adopted. So I need to add the vlan and reconfigure the spanning tree one by one. And it says the procedure should be starting from the root, and just want to know why is it like that.
Thanks very much.
>> But in my environment the VTP method is not adopted. So I need to add the vlan and reconfigure the spanning tree one by one
The answers is that it depends from your current MST configuration : if the new Vlan falls within a range associated to an MST instance and want to use that specific MST instance for the Vlan there is no impact and you just have to create the L2 Vlan on each switch with manual commands because you are not using VTP. However, there will not be any issue because all the key parameters that decide if two switches are in the same MST region have not changed:
MST region name
>>hash of Vlans to MST instances mapping
If the new Vlan is not included in a configured range of Vlan to MST instances it will be associated to instance 0. IF you can accept this, again as explained by Peter Paluch in that thread there is no issue.
Issues arise only if:
you want to add a Vlan that will cause the Vlans sets to MST instances to change because that will change the hash that travel in MST BPDUs and will cause the network to be divided in two different MST regions.
If you need to do this you need to ask for a maintanance time window as there is impact.
You should try to avoid this last scenario.
To achieve this, your MST configuration should be made in a way that it provides for each MST instance a range of Vlans that are associated with, even if these Vlans do not exist. (This is the key point MST does not check / require that the Vlan already exists)
This is an allowed configuration and you can consider it as a sort of provisioning.
The best practice is that if you have 8 instances for example you divide the 4096 Vlan space in 8 subsets and pre-configure the Vlans to MST associations.
Later when you need to create a new Vlan, depending on the desired MST topology, i.e. the selected MST instance, you will pick up a Vlan number within the range associated to the desidered MST.
This is the big change in approach with MST in comparison to PVST+ and Rapid PVST it is not so plug and play and requires thinking ahead = pre provisioning of range of Vlans to MST instances mapping.
If your network scenario is simple and you can accept that the new Vlan will follow the instance 0 topology you are fine. But in a complex network scenario the provisioning approach is the recommended one.
Hope to help