question on STP

Unanswered Question

Hi !


I currently studing for my CCNP certification .

Questions on STP :


is PVST only supperted on ISL ? what about new switch on which ISL is no more supported ?


I'm not sure about the process about the compatibility issue between 802.1d and 802.1w detection protocol, timers....


When I use Extended VLAN should I use STP Extended ID ?


What's meanning when in documentation they are talking about 64 MACs addresses max in relation of some model of switch and type of STP (extended ID or not) ?


When I want using sender bridge ID as a tie breaker in BPDU in STP negatiation process.


for my studdy I read the following Cisco documentation :

Understanding Rapid Spanning Tree

Protocol (802.1w)

Understanding and configuring STP

Understanding and configuring Backbone Fast on Catalyst Switches doc ID 12014

And finally BCMSN official document from Cisco Press.


Excuse me it maybe my English, I'm a French peoples, sometimes English documents are confuse for me I had to aks someone to explain to me in others words...

  • 1
  • 2
  • 3
  • 4
  • 5
Overall Rating: 0 (0 ratings)
Loading.
shrikar.dange Sun, 01/27/2008 - 21:18
User Badges:
  • Bronze, 100 points or more

hi,


1)NO. in a 802.1q (cisco network) the switches maitain one instance of spanning tree for each vlan allowed on the trunk.


2)I am not clear on this statement. What do you are asking I am unable to understand.


3)NO. A Catalyst switch can be configured to use one of the following formats for its STP Bridge ID:

Traditional 802.1D bridge priority value (16 bits), followed by the unique switch MAC

address for the VLAN

The 802.1t extended system ID (4-bit priority multiplier, plus a 12-bit VLAN ID), followed by a nonunique switch MAC address for the VLAN.

If the switch can't support 1,024 unique MAC addresses for its own use, the extended system ID is always enabled by default. Otherwise, the traditional method is enabled by default.

In the extended system-id mode, the default priority is 32,768 plus the VLAN number.And the pririty should be multiple of 4096.


4)When you are recieving BPDUs advertised by root on two different ports from two different switches you compare the sebder BID and the port recieving BPDUS from lowest sender BID becomes fwding.

The criteria for applyin STP is:


1) Lowest root BID

2) Lowest path cost to root bridge

3) Lowest sender BID

4) Lowest port priority

5 Lowest port id.


5)The catalyst switches typically have a pool of 1024 MAC addresses.This pool acts as the MAC address component of BIDs for vlan spanning trees (PVST +).The number of MAC addresses that are availabe depends on swich model.All switches do not support 1024 mac address.


HTH,


regards,


shri :)

2)I'm talking about the manner of the 802.1w works with legacy switch 802.1d ? I was understand the 802.1w switch wait 3 sec and next BPDU is receive on his port it assume it is the protocol have to be use on this port. In case of 802.1d protocol was detected it start to use 802.1d BPDU on this port. One thing it not clear is why if the legacy switch start to use 802.1w BPDU why the first need to be reset his detection.


3a) I understant Extended vlan ID is not related to using STP Extended ID


3b)You say :"If the switch can't support 1,024 unique MAC addresses for its own use, the extended system ID is always enabled by default. Otherwise, the traditional method is enabled by default." what are thoses 1024 MAC address exactly (I was in the impression I had ask that question but it's not there in my text) I whould to understand what is the relation of the number MAC addresses supported and the STP system I have to use... For exemple, if I have 2960 Catalyst switch I know STP Extended System ID is enable by default, but if I use 2960 switch as access layer switch and on distribution I have bigger root switch which can support more that 1024 MAC addresses should I continue to use STP Extended ID ?



Probably here you explain my previous question but it's not clear for me....: (The catalyst switches typically have a pool of 1024 MAC addresses.This pool acts as the MAC address component of BIDs for vlan spanning trees (PVST +).The number of MAC addresses that are availabe depends on swich model.All switches do not support 1024 mac address.

) Are you talking about the number of host the switch can learn location ?? or you talking about the internal 18 MAC addresses I see in 2960 with mention CPU and static ?


thanks a lot for your help it's really appreciate !!



shrikar.dange Tue, 01/29/2008 - 20:44
User Badges:
  • Bronze, 100 points or more

hi,


2)Each port maintains a state that defines the protocol to run on that particular port(eighther .1w or .1D)If the port consistantly keeps receiving BPDUs that do not understand current operating mode for two times the hello time it switches to other STP mode.


3b)For switches that have fewer mac address than number of supported vlans (in your case 2960) in the pririty field of BID first 4 bits are used for priority and last 12 bits are used as system id extension (which the vlan numer always).the priority becomes multiple of 4096 (as only first 4 bits are used) plus the vlan number (system id extension).So the BID becomes unique although you are using same mac for different STP instances.


Yes you are right there.Those 1024 MAC addresses are internal to the switch mentioned as cpua nd static.They are not learned MAC but hard-coded macs.



HTH,



regards,



shri :)

hi !


2) If on switch "A" which run 802.1w, on port 1 I have a switch "B" which use 802.1d, and I changed STP type for 802.1w on switch "B", what append on port 1 on switch "A" is it start to using 802.1w on this port or I have to do some thing ! How this process work exactly ?


3b) OK but, I see thoses MAC addresses exist also on other 2960 switch with exact same values. If those MAC are use to identify the root bridge for a specific VLANs how this bridge can identify if same MAC exists on all 2960 Catalyst switch ? (I hac compare thoses MAC on 2 differents 2960 switchs, and the exacts same lists exist on each of them, with 2 MACs are follow each others)


Second thing if I connect my 2960 Switch with bigger switch like a 6500 Catalyst should I continue to using STP Extended ID (in clear when I have to use STP Extended ID and when I should not ?) I know this is maybe more in than CCNP Examen aks, but I think I have to understand how this is work to be able to decide when I have to use STP Extended ID and when I should not !


thanks a lot !!

hi !


(I think I had badly post my answer today, I post it again and I add some question about RSTP)


hi !


2) If on switch "A" which run 802.1w, on port 1 I have a switch "B" which use 802.1d, and I changed STP type for 802.1w on switch "B", what append on port 1 on switch "A" is it start to using 802.1w on this port or I have to do some thing ! How this process work exactly ?


3b) OK but, I see thoses MAC addresses exist also on other 2960 switch with exact same values. If those MAC are use to identify the root bridge for a specific VLANs how this bridge can identify if same MAC exists on all 2960 Catalyst switch ? (I hac compare thoses MAC on 2 differents 2960 switchs, and the exacts same lists exist on each of them, with 2 MACs are follow each others)


Second thing if I connect my 2960 Switch with bigger switch like a 6500 Catalyst should I continue to using STP Extended ID (in clear when I have to use STP Extended ID and when I should not ?) I know this is maybe more in than CCNP Examen aks, but I think I have to understand how this is work to be able to decide when I have to use STP Extended ID and when I should not !


questions about RSTP :


1) I not sure I'm correctly understand the difference between alternate and backup port. When I had read on it I had suppose wasd correctly understood, but in lab (I use 4 switchs 2 2950 and 2 2960 connected each others) and none port as mount as backup role all blocking was alternated.


2) Was I understand correctly ? The root RSTP network flood BPDU on his designated port at regular basic (HELLO-TIME) is non-root switch received BPDU from root it relayed this BPDU but if it not received BPDU from root it generate BPDU and this one ?? what this BPDU contain (information about the root it self for example ?) If it is what this switch is made I suppose the switch it not send thoose BPDU on his root port ?


3) Is exist some differences between TCN BPDU, TC BPDU and TCA BPDU those appear the sames in some documents for me and differents in others... I one document I was understant 802.1d is able to understabt 802.1w TC/TCA BPDU but 802.1w send also TCN BPDU for compatibility issue with 802.1d....


4) I understand a ROOT bridge send a BPDU evrery hello time on his all designated port as a keep alive messages, when connected designated bridge receive those BPDU they are relayed those on there designated port an so on... When I switch miss 3 BPDU in row that connection as consider lost and the relection process can restart... What append when designated switch did not received BPDU on his root port is it generate BPDU it self and send it on his designated port as a keep alive message ? What is avertising in thoses kind of BPDU as root bridge for exemple ??


5) In case of indirect link failiure should I suppose to able to view RLQ packets pass between siwtches with some sniffer software? I was connected between them on a hub I was see STP packet to orphelin switch advertsign it self as a new root but nothing about RLQ (I was connected on hub between each switch with my sniffer software...)


6) What system are use to notifying Topology Change and what is different when a RSTP switch is connected with STP legacy switch ? When the TC timer is use and is not. I my documentation some time they are talk about TCN BPDU for legacy switch and TC BPDU is it the same thing or not what is the difference ? And if they are different why 802.1d understand


thanks a lot !!



shrikar.dange Thu, 01/31/2008 - 01:56
User Badges:
  • Bronze, 100 points or more

hi,


STP:

2) There is a variable enabled on every port of the switch that defines the protocol to run on the segment.If the port consistantly keeps receiving BPDUs that do not correspond to its current operating mode for 2 times the hello time, it switches to other STP mode.(This is from cisco press book ;) :) )


3)Yes you are right that particular series switches may have same mac addresses.But these addreses are used sequentially.1st mac from address pool will be given to vlan 1 2nd mac to vlan 2 and so on.When all macs are assigned to vlans and there are still some vlans present then this extended sys id will be used(with mac reduction feature enabled.) so you will get unique BID again.One more thing these mac addreses used for STP process for L2 communication bridge uses a single mac (depending on the module).In stp whole BID(priority+ mac) is considered not only a mac address.Use extended sys id only when you have more vlans than supported mac address on the platform.


RSTP:


1) a. alternate port: when your root port fails this port immediately takes over as a root port.In other words this is the "NEXT BEST" root port.

b. Backup port: If you have 2 ports on the same bridge connected to the same segment then the one of the port becomes the backup port.Back up port is the "NEXT BEST" designated port for that particular segment.

It depends on your topology why you do not see alternate and back up port.Its not mandatory they will show up each time you enable RSTP.


2)Yes you are partly correct here.The RSTP bridge relays the BPDUs on all DESIGNATED ports not on root ports.The bridge will send the BPDUs with current information (current root BID) for 3 hello times if they do not recive BPDU from root or neighbouring bridge after that they will clear all protocol information try to converge as fast as possible.

3)TCN bpdu is used only in 802.1D.The TC and TCA bpdus are the configuration bpdus with different bits used for differnt purposes in the flag field of the BPDU frame format.

In 802.1D there are two different type pf message formats of the bpdus as 00(configuration) and 80(TCN) which is not the case with 802.1w wchich uses only 1 msg format 00.so 802.1W sends TCN bpdu for compability issue to its 802.1d neighbour.

4)same as ans 2

5)I did not understand your question properly.But with sniffer you should be able to see the RLQ frames.It depends on your topology and configuration.

6)For TC while timer the "ANY" RSTP bridge which notices the topology chagne sends configuration bpdus with TC flag bit set.In case of STP this is done "ONLY" by root bridge.

TCN bpdu is only used in 802.1D it is different message type all together (80) while TC and TCA are the configuration message types (00) with different bit set in the flag field of the configuration BPDU.

In case you are connecting RSTP bridge to legacy or 802.1D bridge then 802.1w will send TCN bpdu to this legacy switch as on that port both the switches will be running 802.1D

huh!!!!!!! it was a looooooong post i have ever typed!!!!!!



HTH,


regards,


shri :)



Hi


You wrote :


But these addreses are used sequentially.1st mac from address pool will be given to vlan 1 2nd mac to vlan 2 and so on.When all macs are assigned to vlans and there are still some vlans present then this extended sys id will be used(with mac reduction feature enabled.) so you will get unique BID again.One more thing these mac addreses used for STP process for L2 communication bridge uses a single mac (depending on the module).In stp whole BID(priority+ mac) is considered not only a mac address.Use extended sys id only ...


OK I missed something :

I understand if I have 2960 switches and I use 10 VLANs in my network I have to use STP wit not extended ID enable, but if I still use 2960 switches and in my network I have 20 VLANs I have to use STP with Extended ID enable. Is it the same thing if I'm sure I know I will not use more than 10 VLANs per switches ? Which one should I choose STP Extended ID or not ?


If I have more powerfull switches on my network should I use STP Extended ID as well even my more powerfull switches will be my root and secondary root switches ?


I don't undertsand why I have to map VLANs to statics MACs when I has read your answer I was understand thoses MAC was used to identifier ROOT bridge for that VLAN but I find than answer a bad understand from your explication and I had setup a small lab to make sure…


I understand we are using VLAN_PRIORYTY+VLAN_ID+MAC_address_of_the_switch as a ROOT Bridge ID but where is the relation with static CPU MAC address ?


What is the command I can use to see the association between static MAC CPU address I can typed on 2960 switches this will maybe help my to understand some thing…. Should I see same association on eachs switches ?


RSTP Question :


1) For backup it what I had understand when I had read it in documentation, to this messages I and attached a small design what I made in lab (I was maybe used different interfaces but final topology is the same here…. I was understand one port have to go in alternated OK and the second one will be going in backup port. In event of if one of those port need to go in forwarding state the alternated port will be and backup port will the new alternate port. But here both are in alternated state none in backup state (I has left all cost in there default state and all links here are 100mbps full duplex connection) see pdf file : simple RSTP network backup port.pdf


2) Perfect ! it what I had supposed but I was'nt sure !!


3) OK


4) Oups !!


5) I had understand Backbone fast feature is in part of 802.1w and enable by default. But I had forgot some thing when I had asked that question : maybe in 802.1w this feature is not use same appellation RLQ…..


But my question was :


I had setup a small network with always my 4 switches (2 2960 and 2 2950) configure RSTP on them and loop them. See pdf file (rstp indirect link failure.pdf)


If I disconnect the cable on port 2 on hub A my SWD will still have his interface FA0/24 but will no longer BPDU from root I will in case of indirect link failiure.


I had understand : after max age timer expire because SWD did not received no more BPDU from root (SWA) is start as advertising it self as the new root for the network. (for this I had see it when I was connected on the « TEST HUB C ») when SWC received BPDU from SWD advertising it self as the new root, the SWC switch send RLQ to SWD to make sure it has lost connection to SWA. (I did not see this with sniffer software) I had restart all process with my laptop connected on « TEST HUB D » after the SWC switch know the SWD switch as lost connection to SWA, SWC send RLQ to SWB to make sure this switch as connection to SWA because SWB have connection the RLQ relayed RLQ to SWA which reponse positive ACK to RLQ to SWC. (for the complete second part I was not able to observed RLQ between switches when I was connected on « TEST HUB D »)




Jon Marshall Sun, 01/27/2008 - 23:35
User Badges:
  • Super Blue, 32500 points or more
  • Hall of Fame,

    Founding Member

  • Cisco Designated VIP,

    2017 LAN, WAN

Hi


Just to add to Shri's excellent post.


PVST is supported on ISL only. Each vlan runs it's own instance of spanning-tree.


When Cisco needed to run PVST on 802.1q which uses a common instance of STP they created PVST+.


So


PVST = ISL

PVST+ = 802.1q


Jon

shrikar.dange Mon, 01/28/2008 - 01:08
User Badges:
  • Bronze, 100 points or more

hi jon,

Thanx, I miss that point.yes pvst+ is created to communicate with cst & pvst is supported only on isl.

Plus the default mode for catalyst switches is pvst+.


regards,


shri :)

Actions

This Discussion