RSTP

[Phil Storey_1]

Hello to you all, I'm new to the forums, and any help you can give is appreciated.

I have just set up a LAN across 3 buildings on one site. All switches are connected via fibre

In Building 1 Rack A: one 2524 with 2 fibre trancievers
Building 1 Rack B : one 2524 with 2 fibre

Building 1 Rack C : one 2524 with 2 fibre
In Building 2 (GO) there is a stack of 3 2524 switches and two fiber trancievers the switches are stacked together with the GBit stacking modules

Building 2 (GI) there is one 2524 with two Fibre tranceivers

Building 3 Rack A: one 2524 with 2 fibre tranceivers

Building 3 Rack B: one 2524 with 2 fibre
tranceivers

Building 3 Rack C: one 2524 with 2 fibre
tranceivers

Building 3 Rack D: one 2524 with 2 fibre
tranceivers

Building 3 Rack E: one 2524 with 2 fibre
tranceivers
I I have set all switches to use RSTP,

The Stack of 3 units in Building 2 is the primary stack.

When I look at the spanning-tree information, it show the ports 25 / 26 as forwarding on all the switches, Is this correct or should one be forwarding and one blocking or listening ?


I would like to ensure that the spanning tree operates quickly in the event of a link failure

[Shadow13]

kindly post ur network diagram,
also about stacking, in HP stacking is only for telnet not for managing and confguring all of the switches together at the same time

check which spanning-tree version is running.

Regards

[Pablo Rudaeff]

Check the following things:
1. If there are loops in your configuration?
If the answer is no, all the ports in "show
spanning-tree" must be in forwarding or disable
If the answer is yes, you need to manage your tree by giving parameters for the relevant switches (define a root bridge and so, by giving priorities and costs). After that check what ports are in block mode.
Good luck, Pablo.-

[Phil Storey_1]

All Switches are running RSTP

the commander in U2 has the two cascade modules in, the other two are members and have a cascade module and 1 fibre 100BaseFX module in

all other 2524 are running RSTP and all have 2 fibre 100BaseFX modules

when I look at < show spanning-tree > it show the following

Stack Commander

Protocol Version : RSTP
STP Enabled : Yes
Force Version : RSTP-operation

25 1000Stk 20000 128 Forwarding
26 1000Stk 20000 128 Forwarding

Spanning Tree config Operation of commander

Protocol Version : RSTP
STP Enabled [No] : Yes
Force Version [RSTP-operation] RSTP-operation
Switch Priority [8] : 8 Hello Time [2] : 2
Max Age [20] : 20 Forward Delay [15] : 15


Member 1
Protocol Version : RSTP
STP Enabled : Yes
Force Version : RSTP-operation

25 100FX 200000 128 Forwarding |
26 1000Stk 20000 128 Forwarding |

Member 2
Protocol Version : RSTP
STP Enabled : Yes
Force Version : RSTP-operation

25 100FX 200000 128 Forwarding |
26 1000Stk 20000 128 Forwarding |

Other 2524 are single units with no cascade, just x 2 100BaseFX modules

CABxxx

Protocol Version : RSTP
STP Enabled : Yes
Force Version : RSTP-operation

25 100FX 200000 128 Forwarding
26 100FX 200000 128 Forwarding

This is what is on the other switches also.

Protocol Version : RSTP
STP Enabled [No] : Yes
Force Version [RSTP-operation] : RSTP-operation
Switch Priority [8] : 8 Hello Time [2] : 2
Max Age [20] : 20 Forward Delay [15] : 15

It would seem that all units have the same priority ( 8 ) the stack needs to have the highest priority. If a assign a higher priority to the stack is it at port level ?

Do you sign the remaining switch/s - the same level of priority but lower than the commander ?
will the 2524 need to have a warm reset or is the change done on the fly ?

I have attached a drawing but you will need visio 2007 viewer

[Richard Brodie_1]

It probably isn't that important in an ring where the root switch is.

What you should see is just one of the links non-forwarding. If you look at what the root switch is, or configure one manually, the non-forwarding port will be at the opposite side of the ring.

If you wander round the ring of switches the root path cost will go up as you head away from the root switch then start coming down again.

[Phil Storey_1]

If I have 3 2524 units in a stack with the middle unit has two stacking modules and the one above with a 100baseFX nodule in and a stacking module, the bottom unit also has a 100BaseFX module and a stacking module in.

if the middle unit fails I take it the stack has failed, should there be a copper link between the top unit and the bottom unit ?

I know on the Nortel kit you casscde all the units with the final connection being from the last unit to the first unit.

Also with the 3 units in the stack, which one do I give the highest priority ?
is it just the commander ? or do you apply it to each switch working with the commander
with lowest number 1 and the each of the other switches in the stack at 2 and 3 ? but leave all the other switches at 8.

[Pablo Rudaeff]

Check every switch (not stack). You works with the default, I.E. one of the 100MB must be in blocking mode (I assume you linked the stack with 1G).

[Phil Storey_1]

All switches except the Stack are at priority 8, the stack commander is priority 0

All other switches are showing forwarding on ports 25 and 26 .

the stack commander and the two members are connected via the Gigabit stacking module.

If I disable one of the fibre ports on the stack, then re-enable it should it go to blocking ?

[Pablo Rudaeff]

When you will disable the link, in act, you will broke the loop, then every port will be in forwarding. After enable it again, it will be as before, I.E. blocking