Options

Port status on two directly connected 2950 switches

josunin75josunin75 Member Posts: 44 ■■□□□□□□□□
Hello, Cisco Experts, Sirs.

I'd attached two 2950 switches using patch cables, and booted them both up.
The cables are connected as follow:

<Switch1> FA0/1 -> <Switch2> Fa0/1
<Switch2> FA0/2 -> <Switch2> Fa0/2

Initially as it was booting up, I saw the status light changed to green on both switches
But when they finished booting process, the status light on both switches was off.

If I connect PC to any ports on either switch, I see the status light turns into green.

I wanted to study spannding-tree with these switches.
Is this default behaivour? Isn't one port supposed to turn green and the other amber
for one and the other both green?

Is it the cable issue? Do I need to use the cross-over cable? instead of straight through patch cable?

I'd done "write erase" and "delete flash:vlan.dat" and reloaded without saving anything
but I still see no light on the status light of the port.

Any Idea?
=============================
on to BCMSN ^=============^

<Lab>
2 x 3550 EMI
2 x 2950
2 x 3640
1 x 2621
=============================

Comments

  • Options
    kryollakryolla Member Posts: 785
    do sh int trunk and sh spanning-tree to see if both switches are seeing each other if not than you need a cross over cable. Or just do sh int status to see if they are up
    Studying for CCIE and drinking Home Brew
  • Options
    dynamikdynamik Banned Posts: 12,312 ■■■■■■■■■□
    You're supposed to use crossovers for like devices. Some switches automatically detect the type of cable and adjust for it. Do the 2950s do this? You should probably use crossovers just to be sure.

    Also, doesn't the light only refer to the physical connection? i.e. it will be green as long as the connection is good, regardless of the state STP puts it in.
  • Options
    malcyboodmalcybood Member Posts: 900 ■■■□□□□□□□
    yeah I'll back up dynamik, you need a crossover cable. Doesn't matter if the ports are trunked or not, if the light is off the port is either disabled / shutdown or you need to use a crossover cable.

    This is backed up by the fact if you plug a PC in with a straight through RJ45 you see a green light which indicates physical connection.

    If the ports were incorrectly configured for trunking you would still get a green light on the switch port you would just get trunk mismatch errors appearing at the CLI.
  • Options
    Mrock4Mrock4 Banned Posts: 2,359 ■■■■■■■■□□
    Definitely get a crossover. This does, however, sound like the default behavior of a switch that is plugged into another, with one of them configured as a trunk port, and the other as a regular switchport...spanning-tree will run, then put the port into blocking mode, which can be seen in "show int f0/1" or whatever port. I had a guy install a switch that I went out to troubleshoot, with symptoms similar, of it going green, then going back down.
  • Options
    jamesp1983jamesp1983 Member Posts: 2,475 ■■■■□□□□□□
    dynamik wrote:
    You're supposed to use crossovers for like devices. Some switches automatically detect the type of cable and adjust for it. Do the 2950s do this? You should probably use crossovers just to be sure.

    i bet crossovers would remedy this issue. to connect similar layer devices use crossovers.
    "Check both the destination and return path when a route fails." "Switches create a network. Routers connect networks."
  • Options
    Mrock4Mrock4 Banned Posts: 2,359 ■■■■■■■■□□
    Did some double checking, because I honestly didn't remember, but 2950's are *not* Auto-MDIX. It's a safe bet to use the right cable anyways. I've heard of Auto-MDIX jacking up link speed/negotiation..I don't know how true that is, but it's better to be safe than sorry.
  • Options
    kryollakryolla Member Posts: 785
    auto-mdix but I don't think 2950 supports it but really don't rely on lights and use some CLI to verify what your seeing is correct. I hate it when the NOC calls me and asks what lights do I see on said equipment
    Studying for CCIE and drinking Home Brew
  • Options
    josunin75josunin75 Member Posts: 44 ■■□□□□□□□□
    Thanks for the replies, guys...

    Yes, it was the cable. As soon as I used the cross over cables, they both turned green.

    BTW when I did "show spannding-tree", one port was forwarding and the other blocking...

    but doesn't that blocking port's status light should turn amber?

    Both of my FWD and BLK ports are green!

    Is this normal?

    -Thomas
    =============================
    on to BCMSN ^=============^

    <Lab>
    2 x 3550 EMI
    2 x 2950
    2 x 3640
    1 x 2621
    =============================
  • Options
    dynamikdynamik Banned Posts: 12,312 ■■■■■■■■■□
    josunin75 wrote:
    but doesn't that blocking port's status light should turn amber?

    Both of my FWD and BLK ports are green!

    Is this normal?

    This is what I was talking about earlier. I believe the lights are only concerned with layer-1, not layer-2.
  • Options
    Mrock4Mrock4 Banned Posts: 2,359 ■■■■■■■■□□
    I have seen the 2960's when trunking is not negoiated flash green/orange/green/orange etc. Kind of strange, I don't recall seeing that on the 2950's before, but maybe I just didn't pay enough attention (or maybe I am slipping).

    I'm accustomed to a port light going out once it is in the blocking state, so I'm not sure. Maybe someone else will chime in here with some insight.
  • Options
    cisco_troopercisco_trooper Member Posts: 1,441 ■■■■□□□□□□
    I have a double full mesh configuration on four 3750s operating in layer 2 mode with spanning-tree enabled and all lights are green....

    You guys may be thinking about port-security err-disable status, when the interface gets actually shut down.

    In spanning tree, a port in blocking stating is still listening to BPDUs to stay aware of the layer two topology, so it is still active, it just isn't forwarding traffic.
  • Options
    Mrock4Mrock4 Banned Posts: 2,359 ■■■■■■■■□□
    Double full mesh blah blah.....

    Actually, to be fair, I think I am thinking of err-disable.
Sign In or Register to comment.