Regarding PVST & RPVST - sh running-config | section spanning

[13jobsp90]

I have a situation in my office that is - all the cisco switches (2960 model) are running in PVST currently in my network.So if I want to change to RPVST, then globally I have to change accordingly put the command, along with that for each access ports (where hosts are connected) what command should I put other than switchport access vlan command?Can u pls help me?

[MHM Cisco World]

From pvst to r-pvst you need only below command in global 

Spanning tree mode rapid-pvst 

That it.

MHM

[13jobsp90]

When I did for 1 edge switch after changing to RPVST, some of the neighbouring switches vlan went down completely.

[MHM Cisco World]

Can you share 

Show spanning tree 

After you run r-pvst

MHM

[paul driver]

Hello
Suggest for minimal disruption I would start at the access layers switches, changing them to RPVST working your way upstream towards the core switches changing them last to RPVST

[13jobsp90]

When I did for 1 edge switch after changing to RPVST, some of the neighbouring switches vlan went down completely. 

[paul driver]

Hello

---

@13jobsp90 wrote:
When I did for 1 edge switch after changing to RPVST, some of the neighbouring switches vlan went down completely. 

---

Sorry to hear that, tbh you should not see much of an outage on your end hosts, if they they are set up correctly. maybe you can share a topology diagram of your switch infrastructure?

Did those other switches that lost connectivity re-establish after a short period?, If the switch(s) that you changed to rapid-pvst had multiple other switches connected then stp events would be expected however you should not incur outage as such you've lost connection?

The migration needs to be applied on very edge switches first, meaning the only other connection from those switches would be an upstream switch(S) like distribution and ideally all stp edge ports set for fast forwarding( portfast)

Once this has been completed, the you move upstream towards your distribution switches then lastly onto your core switches.
note when you change the cores switches, If any FHRP is applied you may see this flap as/when the stp mode is changed. but you should lost end host connectivity>

Pre-change would expect to have the below in place:
All stp edge ports enabled for portfast (access switches)
All switch interconnects set to stp link port type point-to-point
Stp bridge priority's set in a hierarchical manner ( core<distribution<access-layer)

[13jobsp90]

connected several switches 2 vlans got offline.i can see 2 reasons from my knowledge.

1) edge switch has loopback guard disabled

2) Core switch - when i checked the command - sh running-config | section spanning

there were several vlans showing up, but the vlans which went offline were not shown here. Is that reason for vlans outage?

[Joseph W. Doherty]

Most likely issue for your outage was a re-convergence of PVST, which can take close to a minute.  (That's one of the issues rapid-PVST tries to mitigate.  At best, though, even rapid-PVST is slower than most dynamic routing protocols; one of the reasons L3 designs are recommended over L2 designs.)

[13jobsp90]

the downtime was about for 10 mins by then i have changed to pvst.

[paul driver]

Hello

---

@13jobsp90 wrote:

connected several switches 2 vlans got offline.i can see 2 reasons from my knowledge.

1) edge switch has loopback guard disabled

2) Core switch - when i checked the command - sh running-config | section spanning
there were several vlans showing up, but the vlans which went offline were not shown here. 

---

Not sure what this means but the cores should be aware of all vlans and  stp must not be disabled for any vlans on any switch, Also due diligence is an necessity before attempting a migration like this (some of the steps ive already outlined) and you defiantly need to understand your own switch topology and how it interconnects.

Are you able to share a topology diagram?

[13jobsp90]

It is a large network infra consists about 140 running switches including core switches?How do i share the topology? can i share it with you?

[Joseph W. Doherty]

140, withen same L2 domain?

Are you aware of STP's "diameter" limitations?  Hopefully you haven't exceeded those.

I'm concerned when you noted STP change caused downtime in excess of 10 minutes, which didn't come back, on-line, without further manual intervention, correct?

[13jobsp90]

I have a doubt when a new vlan is created recently in core switch, will it automatically participate in stp process?