2

I have a issue in my network. In terms of OSI I am OK in the physical layer. My link is OK because I tested connecting a computer on a fiber link, and I was able to surf the Internet, but when I make a connection between two Cisco switches, I just have CDP - there is no propagation of VLANs.

There is something even stranger. I have a successful ping in a way. Example: I have a Switches A and B connected through that link. I can ping from B to A and get an answer, but ping doesn't work if I ping from A to B.

It seems very weird to me. I believe it is a VLAN problem, but I do not understand because I can't successfully ping both ways.

I verified that I have the same configurations on both switch ports. I have switchport mode trunk and the native VLAN is 1. I think that the problem is because of a broadcast storm, but I do not know how to verify that. The other situation that I have in my network is this is a multivendor network. My "core" switch is a 3Com. I have Cisco switches, HP switches and TP-Link switches with RSTP.

Should I have a same version of STP in all switches? Could it be a physical problem? Could I have problems with the fiber link? In the moment when the fail occurred, a person was working near to the rack where the fiber ends. I have two transceivers, and I always test between gigabit ethernet ports.

  • I confirm. The link fiber is OK. The trunk link works. However now I have to discover the source of the problem because it is very weird tha fail that ocurred. I am still worry. Before the trunk link was between one Cisco Switch 2960 and 3Com one. Now it is between two Cisco – dfdomin Dec 22 '15 at 14:43
3

You should really run the same version of STP on all the switches. That means converting the Cisco switches to the same STP version as the other switches. Cisco has many documents about PVST and MST interactions if you search for them, and there is this one: Configuration example to migrate Spanning Tree from PVST+ to MST

| improve this answer | |
0

So it's a common issue with Fiber, especially between switches. What it sounds like you need to do is what I've heard referred to as, "roll the fiber". Basically on one end of the Fiber cable, swap them around. Think of cross over cable regarding copper/Ethernet.

| improve this answer | |
  • The switch link wouldn't come up, much less allow CDP, if the the fiber were flipped the wrong way. The question clearly says that CDP works, so it is not a layer-1 problem. – Ron Maupin Dec 26 '15 at 3:09
0

Unless you are using VTP (assuming you are not since you have a mixed environment), multiple VLANS won't be propagated automatically across trunk links unless you have specifically manually created each vlan on each switch and allowed it to be propagated on each trunk interface (by default Cisco allows all vlans). I'm assuming you have already done this, but just wanted to mention it. In regards to the native vlan 1 connectivity between each switch, those should by default be carried across your dot1q interfaces. In regards to your ping not working from A to B,do you have any vrf's on switch A? If so you will need to make sure you source from the correct VRF when pinging. If there are spanning tree issues, loops, or native vlan mismatches etc, you'll be able to see evidence and clues of this by turning on terminal monitor from your telnet or ssh session on your Cisco switches. If you are already directly consoled to the Cisco switches you won't need to do this, and you should see live logging messages with any stp errors. You can also do show logging, to see past stp error logs. Ron's advice to have same version of STP on all switches is excellent, and will also help to simplify things removing chance of further possible STP complexities/anomalies.

| improve this answer | |
  • I have assumed that I have a disorder in my network in the related with STP. Now I am searching like document information STP of the network. Also now I have awareness about the importance of this protocol. All my Switches Cisco have PVST by default and it works between them, but I have discovered, by example, that my root bridge of vlan is 1 a HP Switche in one of the segments of the network and in other one is the 3Com one. – dfdomin Dec 27 '15 at 12:37
0

If possible, connect/link the switches in a hub and spoke fashion rather than a ring fashion. Each uplink should be configured as a trunk. Be sure the trunk type (802.1q probably) is the same on both ends.

If you have multiple VLANs, which it sounds like you do, route between them on your core/hub device.

STP should not really come in to play much in the hub/spoke topology.

| improve this answer | |

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.