I've been having reachability problems with a few of my remote switches and attempted a CiscoWorks RME IOS upgrade to 1 of them last night.
CiscoWorks copied the image down alright but lost it's connection to the switch before it could finish it's stuff. As a result a few of the files from flash are missing.
My question is if I schedule a reload of the router tonight will it re-establish it's connection to the network? It is currently running OK.
Here is some output from the switch that I think is relevant (ie. vtp status, contents of flash, typical access port config, and trunked port config):
SandalwoodYard#show vtp status
VTP Version : 2
Configuration Revision : 331
Maximum VLANs supported locally : 254
Number of existing VLANs : 90
VTP Operating Mode : Client
VTP Domain Name : cob
VTP Pruning Mode : Disabled
VTP V2 Mode : Disabled
VTP Traps Generation : Enabled
MD5 digest : 0x02 0xE0 0x83 0x3C 0x65 0x25 0x04 0x52
Configuration last modified by 172.30.11.202 at 5-18-06 14:56:43
Directory of flash:/
2 -rwx 342 Jun 14 2006 09:35:15 env_vars
3 -rwx 1593344 Jun 14 2006 03:20:16 c2900xl-c3h2s-mz.120-5.WC15.bin
4 -rwx 5417 Jun 14 2006 09:37:19 config.text
3612672 bytes total (2012160 bytes free)
SandalwoodYard#sh run int Fa0/3
switchport access vlan 211
no cdp enable
SandalwoodYard#sh run int Fa2/4
description Fibre Link to City Hall 2nd Floor Data Centre
switchport trunk encapsulation dot1q
switchport trunk native vlan 99
switchport trunk allowed vlan 1,99,211,1002-1005
switchport mode trunk
Basically, I'm wondering if on reboot if the VLANs will be defined properly (by this VTP client pulling VLAN info from the VTP server) or if without a VLAN.DAT file the switch won't have any VLANs defined after a reboot.
If a reboot will cause me problems, is there anything I can do proactively to rectify this (ie. putting the switch in transparent mode)? Note: This switch is at a remote location so not easily accessible.
Let me know if you need more information.
Make sure that the server have a higher configuration revision that this client. eventhough this is a client, if it has a higher configuration revision and it does not have all the vlan that the server suppose to have, this clietn will overwirte the server's vtp due to the client's higher config revision. an example is as follows:
The configuration revision number of the switch that you inserted was higher than the configuration revision number of the VTP domain. Therefore, your recently introduced switch, with almost no configured VLANs, erased all VLANs through the VTP domain.
This happens whether the switch is a VTP client or a VTP server. A VTP client can erase VLAN information on a VTP server. You can tell that this has happened when many of the ports in your network go into inactive state but continue to be assigned to a nonexistent VLAN.
The following link have different scenarios on VTP problems and how to T/S it:
Thanks for your reponse. The version number on the server is 331 like the client. This vtp client switch is already on the network. I'm aware of the issues with vtp revision numbers ... I'm just unsure if this switch will re-establish it's connection to the network on reboot given it's current config AND the fact it does not have a VLAN.DAT file in flash.
VTP is exchanged when the trunk between the client and the server are established, that's when the vtp on the client gets updated. no trunk no vtp exchange.
Can we do like...
i think you should first of all connect then new switch if you have...and make it server and lets the existing server to send its vlan information to it so you will have the all the same vlan on the existing server to the newly attached server and then remove your old one and just reaload it ...and then again connect to the same network...may it will be help full to you...
rate this post if it helps
Thanks for your response. When you say " that's when the vtp on the client gets updated", am I to assume that this is when the VLAN.DAT file gets created on the client? There was a VLAN.DAT on the vtp client switch and during the partial IOS upgrade by CiscoWorks it was lost. I am assuming that the VLAN.DAT file on the VTP client is needed since it used to be there.
Been watching this thread and have a concern that I have not seen addressed.
It is my impression that the vlan.dat file contains the vtp domain. When I erase the vlan.dat file it set its back to blank as well as vtp server mode.
If you do not use a password I suspect the switch will come up in the correct vtp domain since it seems to set itself to the first it hears but it will be in server mode. You would need to change it back to client. If you use a password it will stay set to blank so it will not get any data and you must manually set the vtp domain and password.
I tested this on 2 of my lab switches and it appears you can get the client to recreate the vlan.dat file by adding a vlan on the vtp server. If you do this it should be ok when you reboot
the amount of aggrevation caused by vtp in a network is now visible...
My question is HOW big is this network? Do you even need vtp ?
If your network is under 20 switches (and it should be even if you have more than 20 switches, as you should be breaking things up with layer 3 routing after 3 switches anyway, just set vtp to transparent (or even better, OFF if you have CATOS).
VTP is not really worth the extra effort, unless you have a really old switched network with many switches (like a college campus) and have not redesigned to a layer 3 routed environemnt from distribution and up yet. IMHO vtp is a dog with fleas for all the reasons we have seen in the thread...(GVRP is even worse)
I have migrated most of the switches to vtp transparent mode but just have not got them all done yet ... so much to do and no one to delegate to :-) ... oh well, keeps me out of trouble!
Fantastic! I just deleted a VLAN I know longer need on my 6513 which is the VTP server and voila ... all of a sudden a VLAN.DAT file appeared on the switch in question.
Thanks for your assistance, Ian.