VTP issue

[Martin Skolar]

Hello,

I am having an issue with VTP on C3850 switch.
VTP server has been replaced an all other switches in the VTP domain work just fine, except one.
This particular switch still points towards the old VTP server and usual procedure for switching between transparent-client, changing VTP domain and password, creating new revision on the VTP server (new VLAN) unfortunately did not help.
Trunk towards the VTP server looks fine as well.

I am getting a suspicion that vlad.dat file might be corrupted.
I was thinking of following procedure:
- vtp mode transparent
- delete vlan.dat
- (maybe? not sure if it is contained in the vlan.dat file) vtp version 3, vtp domain xxx, vtp password yyy
- vtp mode client


Could you please advise if deleting those vlans as per described procedure could cause any issues? My main concern is loosing any VLANs as I have only remote access to the switch, therefore the idea is to avoid any need for on-site intervention.

Do you have any other idea what to try, in order to force the VTP to "recover" and use the new VTP server?

Cheers
Martin

[Martin Skolar]

I did debug of VTP using "debug sw-vlan events"

Jan 16 2023 12:08:19.895 UTC: VTP LOG RUNTIME: VTP3: Mode change to client. Saving outstanding db changes
Jan 16 2023 12:08:20.033 UTC: VTP LOG RUNTIME: VTP3[VLAN]: Db is valid. Flooding db on all trunks
Jan 16 2023 12:08:20.034 UTC: VTP LOG RUNTIME: VTP3[VLAN]: tx vtp summary, domain xxx, rev 116 window 1492
Jan 16 2023 12:08:20.037 UTC: VTP LOG RUNTIME: vtp_pruning_notification
Jan 16 2023 12:08:20.037 UTC: VTP LOG RUNTIME: VTP3 domain xxx modified

It looks the same as my lab environment where the VTP is running fine.
Unfortunately it is not possible not to use VTP here.

[MHM Cisco World]

I need to see the config 

[paul driver · Accepted solution]

Hello
Have you changed anything on the bp server, if so you may need to re-promote the vtp primary server

VTP primary server switch
vtp primary server force

On the client sw
Make it vtp 3 client, Verify the password /domain are the same as the other switches in the estate, then change the vtp ver to 3 this should then initiate a synchronization of the switches vtp DB to that of the new vtp3 primary server vtp D/B 

[Elliot Dierksen]

There is a brute force option. You could copy the vlan.dat file from a working switch (preferably a client switch) to the problem switch. That would overwrite it, and then you would reboot the problem switch. Use that really carefully since you could cause a big problem if you overwrite the wrong switch, not to mention that the reboot will be disruptive to your network!

[Martin Skolar]

Thank you for proposed brute force method, at the moment it is considered as plan "b" for the VTP recovery.
Will let you know how it worked out.

[paul driver]

Hello
Isolate the switch, delete the vlan.dat from flash:
reload switch, after reload check the revision number of switch (should be 0)
change vtp mode to client, reattach switch to production.

[David Ruess]

Hello,

I believe VTP version 3 VLANs are stored in the NVRAM and not the vlan.dat file. Secondly what do you mean the VTP server is still pointing to the "old server" ? Can your switch reach the new VTP server?

You could try changing the VTP version to version 2 and clearing out the domain and switching it to transparent mode and then back to client and then back to version 3. Make sure domain/passwords match and version is 0, however this is less of a concern with VTP version 3 as long as you have the primary server configured that I mention below.

Make sure on your new VTP server you do the following commands:

vtp mode server vlan

Under privileged exec issue the following command:

S1#vtp primary

This system is becoming primary server for feature vlan
No conflicting VTP3 devices found.
Do you want to continue? [confirm]

This should ensure the primary VLAN for all switches in domain

Hope that helps

-David

[Martin Skolar]

Hello all,

In the end of the day the VTPv3 issue was resolved by issuing "vtp primary server" on the VTP server.
It identified the conflicting switch and overwritten it, therefore it resolved the issue.

Thank you all for you help!

[Marc Neoberg]

VLAN Database

!important! en / conf t / ip routing op DC1!

enable

Conf t

Vlan 10

Name xxxx

———————————————————-

(VLAN delete? = No VLAN 49)

(Command cancel Ctrl shift 6)

(Network 24 = 255.255.255.0)

———————————————————-

VTP (DC1= servermode / AS1/2/3 = clientmode)

en

Conf t

Vtp mode server/client

Vtp domain xxxxxxx

Vtp password xxxxxxx

End

Show vtp status <—

Vlan gateways config (Assign VLAN into interfaces

en

Conf t

Interface Vlan 10

Ip address 172.16.10.1 255.255.255.0

Exit

(gebruik gedefinieerde ip tabel)

Bewijs via show run en dan alle ip toekenningen pakken

Trunkpoorten aanbrengen tussen switches

en

Conf t

Interface GigabitEthernet1/1/2

Switchport mode trunk

Exit

(Let op poortindeling tabel - daar staat welke interfaces je op swtichport mode trunk moet zetten)

Spanning Tree Root bridge & Priority