Trunk, STP and duplex

Discussion in 'Cisco' started by newbie, May 24, 2007.

  1. newbie

    newbie Guest

    Hi,

    Obviously, I am missing something very basic.

    I want to hook a new switch (3560) to an existing one (2950).

    On the existing 2950, I define a port as trunk and I define an encapsulation
    mode.
    I define the corresponding port on 3560 with the same parameters.

    I get all the VLan on the new switch and everything is fine.
    I am able to see the trunk port STP status on the 2950 as "forwarding".

    Now the point I do not understand : whenever I put one of the 2 ports (on
    2950 or 3560 end) in half duplex mode, the port on the 2950 will be force to
    "Blocking" state.

    (And actually, I have troubles to get it back even forcing the ports in full
    duplex.)

    Could explain me the root cause for it ? I feel that I am missing an
    important point.

    Thanks in advance,

    A puzzled newbie..
    newbie, May 24, 2007
    #1
    1. Advertising

  2. newbie

    Trendkill Guest

    On May 24, 6:49 pm, "newbie" <> wrote:
    > Hi,
    >
    > Obviously, I am missing something very basic.
    >
    > I want to hook a new switch (3560) to an existing one (2950).
    >
    > On the existing 2950, I define a port as trunk and I define an encapsulation
    > mode.
    > I define the corresponding port on 3560 with the same parameters.
    >
    > I get all the VLan on the new switch and everything is fine.
    > I am able to see the trunk port STP status on the 2950 as "forwarding".
    >
    > Now the point I do not understand : whenever I put one of the 2 ports (on
    > 2950 or 3560 end) in half duplex mode, the port on the 2950 will be force to
    > "Blocking" state.
    >
    > (And actually, I have troubles to get it back even forcing the ports in full
    > duplex.)
    >
    > Could explain me the root cause for it ? I feel that I am missing an
    > important point.
    >
    > Thanks in advance,
    >
    > A puzzled newbie..


    Trunks must be full-duplex, and must be the same speed.

    >From Cisco:


    http://www.cisco.com/en/US/products...figuration_guide_chapter09186a00801fc1e5.html

    In order to assign a port to a trunk, the port must comply with the
    following requirements:

    ·A layer 3 interface is not configured on the port.

    ·A VLAN is not configured on the port.

    ·The port is not assigned to a different trunk.

    ·An available MAC address exists which can be assigned to the port.

    ·Auto-negotiation mode is not configured on the port.

    ·The port is in full-duplex mode.

    ·All ports in a trunk must operate at the same rate.

    ·All ports in a trunk must have the same ingress filtering and tagged
    modes.

    ·All ports in a trunk must have the same back pressure and flow
    control modes.

    ·All ports in a trunk must have the same priority.

    ·All ports in a trunk must have the same transceiver type.

    ·All ports in a trunk must belong to the same module, that means they
    must be located on the same slot.
    Trendkill, May 25, 2007
    #2
    1. Advertising

  3. newbie

    Trendkill Guest

    On May 24, 7:19 pm, Trendkill <> wrote:
    > On May 24, 6:49 pm, "newbie" <> wrote:
    >
    >
    >
    > > Hi,

    >
    > > Obviously, I am missing something very basic.

    >
    > > I want to hook a new switch (3560) to an existing one (2950).

    >
    > > On the existing 2950, I define a port as trunk and I define an encapsulation
    > > mode.
    > > I define the corresponding port on 3560 with the same parameters.

    >
    > > I get all the VLan on the new switch and everything is fine.
    > > I am able to see the trunk port STP status on the 2950 as "forwarding".

    >
    > > Now the point I do not understand : whenever I put one of the 2 ports (on
    > > 2950 or 3560 end) in half duplex mode, the port on the 2950 will be force to
    > > "Blocking" state.

    >
    > > (And actually, I have troubles to get it back even forcing the ports in full
    > > duplex.)

    >
    > > Could explain me the root cause for it ? I feel that I am missing an
    > > important point.

    >
    > > Thanks in advance,

    >
    > > A puzzled newbie..

    >
    > Trunks must be full-duplex, and must be the same speed.
    >
    > >From Cisco:

    >
    > http://www.cisco.com/en/US/products/hw/optical/ps2001/products_config...
    >
    > In order to assign a port to a trunk, the port must comply with the
    > following requirements:
    >
    > ·A layer 3 interface is not configured on the port.
    >
    > ·A VLAN is not configured on the port.
    >
    > ·The port is not assigned to a different trunk.
    >
    > ·An available MAC address exists which can be assigned to the port.
    >
    > ·Auto-negotiation mode is not configured on the port.
    >
    > ·The port is in full-duplex mode.
    >
    > ·All ports in a trunk must operate at the same rate.
    >
    > ·All ports in a trunk must have the same ingress filtering and tagged
    > modes.
    >
    > ·All ports in a trunk must have the same back pressure and flow
    > control modes.
    >
    > ·All ports in a trunk must have the same priority.
    >
    > ·All ports in a trunk must have the same transceiver type.
    >
    > ·All ports in a trunk must belong to the same module, that means they
    > must be located on the same slot.


    Well that link was for link-aggregation, not vlan trunking, but I'm
    pretty sure the rules still apply. I can't decisively say that half-
    duplex wont work (as I have not tried it), but I'm fairly certain that
    full-duplex is a requirement for trunks. As soon as you move to half-
    duplex, the port is taken out of trunking. I'm not sure why it is
    shown as 'blocking', presuming you did a show spantree and that is
    what you are talking about. If the switches only have one
    interconnection, and this downstream switch does not have another link
    to elsewhere in the network, it should be forwarding on all ports in
    all VLANs. If it does have another link, then the blocking may make
    sense depending on root and spantree priorities.
    Trendkill, May 25, 2007
    #3
  4. newbie

    newbie Guest

    Hi,

    "Trendkill"
    >
    > > Hi,

    >
    > > Obviously, I am missing something very basic.

    >
    > > I want to hook a new switch (3560) to an existing one (2950).

    >
    > > On the existing 2950, I define a port as trunk and I define an
    > > encapsulation
    > > mode.
    > > I define the corresponding port on 3560 with the same parameters.

    >
    > > I get all the VLan on the new switch and everything is fine.
    > > I am able to see the trunk port STP status on the 2950 as "forwarding".

    >
    > > Now the point I do not understand : whenever I put one of the 2 ports
    > > (on
    > > 2950 or 3560 end) in half duplex mode, the port on the 2950 will be
    > > force to
    > > "Blocking" state.

    >
    > > (And actually, I have troubles to get it back even forcing the ports in
    > > full
    > > duplex.)

    >
    > > Could explain me the root cause for it ? I feel that I am missing an
    > > important point.

    >
    > > Thanks in advance,

    >
    > > A puzzled newbie..

    >
    > Trunks must be full-duplex, and must be the same speed.
    >


    <<
    Well that link was for link-aggregation, not vlan trunking, but I'm
    pretty sure the rules still apply. I can't decisively say that half-
    duplex wont work (as I have not tried it), but I'm fairly certain that
    >>


    Well ... I prefer to speak of "etherchannel" or "teaming" and avoid "trunk"
    in that meaning.
    No wonder why...

    To make it worse, you find in Cisco documentation paragraphs such as :
    "Configuring EtherChannel and VLAN Trunks"
    I'd rather that they kept the two things clearly separated.

    <<
    full-duplex is a requirement for trunks. As soon as you move to half-
    duplex, the port is taken out of trunking. I'm not sure why it is
    >>


    I checked several books and saw nothing more precise.

    In a BCMSN book, I just stumbled onto that note :
    "Verify that the speed and duplexing configuration on both sides are correct
    and that you're using the correct cable type (crossover vs straight)".
    That's all.

    <<
    shown as 'blocking', presuming you did a show spantree and that is
    what you are talking about. If the switches only have one
    >>


    Yep, I checked the spanning-tree status. On the 2950 sw I have 2 others
    trunks which are forwarding.

    <<
    interconnection, and this downstream switch does not have another link
    to elsewhere in the network, it should be forwarding on all ports in
    all VLANs. If it does have another link, then the blocking may make
    sense depending on root and spantree priorities.
    >>

    The 3650 is a "leaf" switch. It behaves the same even if nothing is hook on
    it.

    The port is blocked as soon as I put it in "half duplex". Otherwise it works
    fine.

    My concern is that I don't see conceptually why it should require full
    duplex.
    I fear to miss something important.

    Regards,
    newbie, May 25, 2007
    #4
    1. Advertising

Want to reply to this thread or ask your own question?

It takes just 2 minutes to sign up (and it's free!). Just click the sign up button to choose a username and then you can ask your own questions on the forum.
Similar Threads
  1. Pete Mainwaring

    Gigastack GBICs and STP hops

    Pete Mainwaring, Jul 11, 2003, in forum: Cisco
    Replies:
    1
    Views:
    1,903
    Chris Marva
    Jul 11, 2003
  2. John Sasso
    Replies:
    2
    Views:
    2,926
    John Sasso
    Aug 14, 2004
  3. STP and VLAN tagging

    , Mar 29, 2005, in forum: Cisco
    Replies:
    2
    Views:
    4,458
    Scooby
    Mar 29, 2005
  4. Imayroam

    STP and Apple Mac's

    Imayroam, Jun 13, 2005, in forum: Cisco
    Replies:
    3
    Views:
    917
    Imayroam
    Jun 20, 2005
  5. DigitalSierra

    ECP Half-duplex or Full-Duplex?

    DigitalSierra, Oct 18, 2004, in forum: A+ Certification
    Replies:
    0
    Views:
    638
    DigitalSierra
    Oct 18, 2004
Loading...

Share This Page