If a VTP client switch is added to the domain, and this client has a revision number that is higher than the current server switch, why does the new switch's configuration get propagated to the server and beyond?
I know how revision number works, and that you should make sure that the server has the highest number before adding a new switch.
But I thought the whole point of server-client relationship was to determine who got to give out new info?
Mitch@_._ wrote in news: snipped-for-privacy@4ax.com:
Clients can't originate new configurations. However, if the newly added switch was a VTP server, it would overwrite the configs on the other switches, including VTP servers with lower revision numbers.
Regards, Lars Christensen CCIE #20292
Thanks.
So when a new client is inserted into a network, it can't mess things up? The server will propagate the VTP info to it?
I found the slide in the curriculum that got me confused.
"There is one VTP server switch, S1, and two VTP client switches, S2 and S3.
S1 / \\ S2 S3 \\ \\ S4 (new switch)
S4, which has been previously configured as a VTP client, is added to the network. The revision number of the switch S4 is 35, which is higher than the revision number of 17 in the existing network. S4 comes preconfigured with two VLANs, 30 and 40, that are not configured in the existing network. The existing network has VLANs 10 and 20.
When switch S4 is connected to switch S3, VTP summary advertisements announce the arrival of a VTP-enabled switch with the highest revision number in the network. The animation shows how switch S3, switch S1, and finally switch S2 all reconfigure themselves to the configuration found in switch S4. As each switch reconfigures itself with VLANs that are not supported in the network, the ports no longer forward traffic from the computers because they are configured with VLANs that no longer exist on the newly reconfigured switches."
Why does this happen, if S4 was a client?
Mitch@_._ wrote in news: snipped-for-privacy@4ax.com:
To sum up the words:
- vtp mode server: Switch originates the VTP messages to all switches with a lower revision number. If other switches in the network is configured as server, the server with the highest revision number wins. If you introduce a new switch into an existing network and it is configured as server, you could therefore end up in a situation, that the newly installed switch actually overwrites the current VLAN database in the network by fault.
- vtp mode transparent: Switch passes the VTP messages through without learning the VTP message. This type is also needed to use extended VLANs (above 1005)
- vtp mode client: Switch will learn VTP messages from any server with a higher revision that the latest VTP database installed. The switch cannot originate VTP messages, but will learn and pass through the messages.
Hope this clarifies your question
Regards, Lars Christensen CCIE #20292
agreed - except the default setting is VTP server (or used to be)....
anywhere with a reasonabley paranoid change control regime will not like the idea of "1 mistake can melt the network" - which is exactly what you can get with VTP.
And now we have the latest WAN technology VLPS (bridging) - so you can actually manage to melt a complete set of campus networks.....
so - the most important VTP setting is "transparent" cos then it is effectively turned off.
I'm curious why there's no direct command to manually reset the revision number? I know there are ways to make it reset, but why no command?
Lars Have you tested your answer in a lab? Do this, or look at this flash,
Mariusz 'BB' Trojanowski wrote in news:gvtmau$erl $ snipped-for-privacy@inews.gazeta.pl:
Hi Mariusz
Watching the flash file, I get a bit confused as the flash file kind of contradict some of the ideas behind the client/server mode in VTP. I haven't tried it the lab, but it might be interesting to do. I'll have to test it before I can conclude on this issue.
Regards Lars
I've got my CCNA exam schedules for next Tuesday. I'M SCARED! :)
So much information. Root bridge. Designated router. VTP. STP. It's like soup in my brain.
I really need to cram extended ACLs. The syntax is very confusing to me.
I started the CCNP track this week, beginning with BCMSN, and it beautifully clarified all things VTP.
Also a very good review for the upcoming CCNA exam.
Setting to transparent mode resets the revision number to zero (amongst other things). This setting will remain if you then switch back to client mode.
FWIW I just went on a CCNA course and the instructor, a guy with plenty of real world experience, said that even when a switch is in client mode, it will propogate its vlan config on joining a network, provided of course it has the highest revision number.
IMO safest thing would be to wipe all config off the switch, then set to transparent mode before adding it to the network.
Yes, I finally came across multiple sources that confirm that, so I feel solid there.
Cabling-Design.com Forums website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.