Hot answers tagged

23

You can't manually force updates because VTP configs are pushed based on the highest current VLAN database revision number. Normally, after every change to your VTP config the configuration revision number should be incremented. The only way to force update this is by actually doing a VTP database change.


13

If VTP is not updating the VLANs correctly, check to make sure the MD5 digest on all VTP domain switches is the same. The MD5 digest is calculated using the revision number, domain name, and password to make sure the information in the VTP update can be trusted. I've seen issues where switches configured with the same domain and password, but the MD5 digest ...


9

In theory, you'll want to migrate the clients to transparent before you remove the server from the equation. This makes absolute certain that things will continue to function without any gotchas. In practice, it doesn't matter. A server will just stop sending summary advertisements to the VTP domain and the clients will continue using the last seen ...


9

Updates to VTP are only sent out when there is an increment to the VTP revision number. The only way to 'force' an update is to create an erroneous Vlan and then delete it. This will accomplish a 'force update' and send the change out to the rest of your clients.


9

This is correct! When Cisco implemented VTP, they left a "null" value for exactly this reason. If a VTP client or server with a null domain receives a VTP message with the domain populated, it will assume the domain of the received message and add applicable VLANs to its database. You can find more information here: Catalyst 6500 Series Switch Cisco IOS ...


9

You can see the primary server (if any) in the output of show vtp status: Primary ID : 0011.2233.4455 Primary Description : primary-switch-name If the Primary ID is 0000.0000.0000, then your current switch does not recognize any primary server, including itself. To answer your questions: When you try to promote a ...


8

Neither Native VLAN or VTP domain mismatch will cause this, as it normally impedes VTP propagation which you obviously do not have a problem with (your 3750 received and applied the VTP information from the 2960) Check for VTP database version. If the 2960 had a higher version than the 3750, then it will be the one used. This is the most likely thing to ...


7

That’s because VTP functions on a client/server model, not a distributed model. The server effectively controls everything that happens inside the VTP domain. When a switches directly connected port/downstream port, it sends a VTP Join message indicating participation in that VLAN. If you wanted something similar to a distributed model that supports VLAN ...


6

VTP pruning is enabled on the server because the server decides if pruning is used in the domain or not. For pruning to be effective it needs to be enabled in the entire domain. If a client was to enable pruning then it would only have effect locally. This would not have any positive effect in the domain except for the device that you had configured it on. ...


5

An alternate way to force an update is to change the switch that needs the update to Transparent mode which resets the counter to Zero and then put it back into client mode or server mode. This will force the switch to jump from revision 0 to whatever the current highest revision is and the database will be overwritten. Hope this helps.


5

Cisco says: You must configure VTP to transparent mode before you can create a private VLAN. Private VLANs are configured in the context of a single switch and cannot have members on other switches. Private VLANs also carry TLVs that are not known to all types of Cisco switches. Reference: Private VLANs Actually if you have a look at the structure of a ...


5

After checking this thread on the Cisco Learning Network and subsequently executing some checks on both the VTP Client and Server, e.g: sh vtp st sh vtp password sh int trunk it appears that the domain is identical and the portchannel is in trunking mode. However, there was a discrepancy between the VTP passwords for some reason. After synchronizing the ...


5

For VTP to work between a VTP server and client, the VTP Domain Name and VTP passwords must match. You have the same VTP Domain Name (ccna), but the VTP Passwords do not match. You can tell that the passwords do not match because the MD5 Digests do not match: Switch 1: 0x44 0xD7 0x99 0x68 0xF8 0x18 0x7F 0xBD 0xA4 0xDE 0xB5 0xC3 0x00 0x69 0xC3 0xC6 Switch 4: ...


4

Not exactly a reason not to use VTP pruning but just a few attention points: If running version 1, transparent mode will disable forwarding of VTP packets, hence making all downstream switches outside of your VTP domain and unaware of VLANs. If you're running version 2, a switch in transparent mode will forward every VTP packet if the switch is set to ...


4

If you have not previously configured a switch to be in a VTP domain, it will become part of the VTP domain from the first VTP advertisement it sees. See the sections which I emphasized below: Configuring VTP: The VTP Domain A VTP domain (also called a VLAN management domain) consists of one switch or several interconnected switches under the same ...


4

I looked again and found I was using the wrong ports. The switches were configured with a GT96100-FE (2 ports) and a NM-16ESW module. I was using the GT9600 ports (fa0/0 -fa0/1) which don't even support switchport mode commands. I plugged everything into the NM module and it worked as expected. Sorry my bad!


3

It seems you have the SVI for VLANs 10, 20, and 30 configured on all three switches. This is going to cause an IP conflict. The only switch that actually needs the SVI is whatever switch that will be doing the routing. All the others just need to simply forward data at L2. From what I can gather, that would be your Switch 3. I would suggest reading through ...


3

VTP only propagates on trunks, so you must have a trunk link between the switches.


3

The short answer is 'no' there is no downside to VTP pruning that I'm aware of. @Pedro Brito gives some good background on the operation and gotcha's of VTP. Since you are using site specific VTP domains the risk of revision number issues is lessened. I would just add to be very specific which switches are your VTP servers at each site and set the rest ...


3

With VTP, you can mark VLANs as eligible to be pruned for a trunk link. When pruning is enabled in VTP, the eligible VLANs will be pruned if there is nothing downstream of the trunk link using those VLANs. This Cisco document shows how to configure VTP pruning and how to add VLANs to the pruning eligible list. switch(config-if)# switchport trunk pruning ...


3

The basic answer to your question is no, the different switch versions don't matter. As long as you're running the same VTP version on the server and client, both are configured for the same VTP domain and password (optional), and the link between switches uses 2 correctly-configured trunk ports, you should be good to go.


3

VTP is a layer 2 protocol and therefore can't be forwarded over a layer 3 boundary. Also, since Vlans terminate at a layer 3 boundary, why would you want to?


3

If a switch gets revision X+1 from one VTP server, it will simply look at the revision number if that same revision number comes in an update from the other VTP server. The switch will then ignore the revision with the same number from the other server, thinking it is up to date. Also, remember that it is not only servers, but also clients, that can ...


3

Neither. DHCP works by the server receiving the client's broadcast - so you need a working broadcast domain (aka L2 segment aka VLAN) in between, unless you're using a DHCP relay. As long as the network is functioning, DHCP will work. xSTP, VTP, or DTP all help establishing and maintaining a working network. There are no 'most preferable' protocols in ...


3

Vtp pruning is used in network for reduced unwanted brocast traffic of vlans . Only choosen vlan brocast traffic will be allowed into choosen links .so that network bandwidth is optimised..and there will be no latency in network.


2

Its true. The 2nd switch will be assigned the domain from 1st switch. A quick test to prove it. Quick output: running-configs do not have the vtp configurations. ! Switch#show run Building configuration... . . . hostname Switch ! ! ! ! spanning-tree mode pvst ! interface FastEthernet0/1 ! . . . ! interface Vlan1 no ip address shutdown ! ! line con 0 ! ...


2

Set vtp mode transparent to disable VTP on a switch. If VTP version 2 is used, a transparent switch will still forward VTP advertisements, but, in any VTP version, the transparent switch will itself ignore VTP. The VTP domain name and password must match between switches for VTP to work, and blank (null) VTP domain names and passwords are valid, so any ...


2

Apparently, your VTP password is incorrect on the 2960X which doesn't work. The MD5 digest on VTP doesn't match on the last (non-working) of the three switches you show. You wrote there is no MD5 checksum error, but there is in the debug output: Apr 25 04:26:06.207: VTP LOG RUNTIME: MD5 digest failing You can use the show vtp password command on a switch ...


2

OK. You have a couple of problems. First, you need to make sure that the channel is a trunk. VTP only works on trunks. On both switches: interface Port-channel1 switchport mode trunk ! Next, you really do not have the same VTP password on both switches. You can tell this because the MD5 hash is different on each switch: Switch 1: MD5 digest ...


Only top voted, non community-wiki answers of a minimum length are eligible