1
00:00:00,760 --> 00:00:06,670
What you'll notice the spending tree convergence is very quick and that's because we're running rapid

2
00:00:06,750 --> 00:00:16,040
previous t plus I didn't make any changes on the switches from a spending tree point of view.

3
00:00:16,110 --> 00:00:24,390
So the command show run piping CLEET span shows the default config and notice the spending tree use

4
00:00:24,390 --> 00:00:30,830
these rapid PV T in the output it shows as always T-P or Reppert rapid spending tree.

5
00:00:30,980 --> 00:00:34,710
But this is actually Rapide per villaine spending tree.

6
00:00:34,910 --> 00:00:38,810
We can change the mode of spending tree.

7
00:00:38,930 --> 00:00:48,070
But before I do that just to make the point again at the moment Sze spending tree on switch three shows

8
00:00:48,100 --> 00:00:52,970
that to the root port is gigabit is 0 1 the support.

9
00:00:53,260 --> 00:00:55,980
If I shut that port down

10
00:00:59,320 --> 00:01:06,250
and then top shows spending tree again what youll notice is gigabit to 0 0 is the root port.

11
00:01:07,570 --> 00:01:15,660
Convergence is very quick with rapid spending tree because it doesnt use timers.

12
00:01:15,880 --> 00:01:21,010
So the max age and for delay time is not to be used for convergence.

13
00:01:21,130 --> 00:01:27,020
The switches send messages to each other with rapid spending tree to enable quick convergence.

14
00:01:27,040 --> 00:01:35,650
So now if we change the spending tree type mode to Peavey's t we should see that spending tree takes

15
00:01:35,650 --> 00:01:39,020
a lot longer to converge.

16
00:01:39,100 --> 00:01:42,280
So I'll change that on the old switches.

17
00:01:44,830 --> 00:01:48,760
So which three switch for

18
00:01:55,250 --> 00:01:56,030
switch five

19
00:02:00,210 --> 00:02:04,450
subaccount switch three show spanning tree.

20
00:02:04,830 --> 00:02:11,340
Notice we can see that the switch is still learning which ports are the root port designate a port or

21
00:02:11,340 --> 00:02:14,790
blocking port when in the learning state.

22
00:02:16,080 --> 00:02:23,730
Traffic will be blocked user traffic will only be forwarded when ports transition to a forwarding state.

23
00:02:23,730 --> 00:02:28,910
At the moment you can see that the spending tree protocol used or shown here is trouble.

24
00:02:29,190 --> 00:02:36,060
But once again you need to be careful because on Cisco switches even though it displays on trouble we

25
00:02:36,100 --> 00:02:42,300
are actually using previous t Peavey's t is once again backward compatible so it will be able to talk

26
00:02:42,330 --> 00:02:43,030
to an ADA.

27
00:02:43,160 --> 00:02:50,210
Toward D switch from another vendor as an example so we can see I triple E in the output here.

28
00:02:51,840 --> 00:02:55,560
So once again shows spanning tree the root port.

29
00:02:55,560 --> 00:02:57,180
In this case is now gigabit.

30
00:02:57,200 --> 00:03:00,890
0 0.

31
00:03:00,990 --> 00:03:08,310
So what happened because previously we had to switch one as the root.

32
00:03:08,430 --> 00:03:09,570
It still has a command

33
00:03:12,420 --> 00:03:20,140
show spending tree shows us that the switch is the root of the topology but get 0 or 1 is not shown

34
00:03:20,140 --> 00:03:21,070
in the output.

35
00:03:22,170 --> 00:03:28,100
Because I need to know shut to that port so can t interface gigabit to 0 or 1.

36
00:03:28,170 --> 00:03:40,030
No Schutt shows spending tree noticed that port is a listening port gig but 00 is a blocking port so

37
00:03:40,030 --> 00:03:42,640
this port is blocking this port is listening

38
00:03:45,460 --> 00:03:51,850
and what you'll notice is it'll take it a while to converge or put an IP address on the switch.

39
00:03:53,120 --> 00:03:53,990
We'll be waiting.

40
00:03:54,020 --> 00:03:55,500
And then I'll demonstrate this again

41
00:04:01,030 --> 00:04:03,280
put my address on which one

42
00:04:14,300 --> 00:04:25,450
will know shut the interface so paying ten on one one to one can paying itself switch three.

43
00:04:25,590 --> 00:04:27,770
No shut the interface.

44
00:04:27,900 --> 00:04:33,410
Paying 10 to 1 to 1 to 1 pink succeed so I'll just do that again.

45
00:04:33,720 --> 00:04:37,350
Notice the ping from switch three to switch one succeeds.

46
00:04:37,770 --> 00:04:46,760
Show spanning tree convergence has taken place because gigabit 0 1 is the root port and is forwarding.

47
00:04:46,950 --> 00:04:57,970
But now if I shutdown gigabit 0 1 and then try paying switch 1 The port has gone down but pings are

48
00:04:57,970 --> 00:05:05,520
failing even though we have a redundant link show spanning tree shows me that of Route port is still

49
00:05:05,520 --> 00:05:16,190
learning still learning pings fail previous t takes a long time to converge.

50
00:05:16,190 --> 00:05:19,420
Can take 30 seconds for that convergence to take place.

51
00:05:19,700 --> 00:05:26,360
As you can see there it's just happened show spending trees shows us now that kickabout 0 1 is forwarding

52
00:05:28,550 --> 00:05:38,400
But once again if I know Shutt gigabit 0 1 and did the ping again the ping would fail because it now

53
00:05:38,400 --> 00:05:41,340
needs to learn that this is the better path.

54
00:05:41,460 --> 00:05:42,660
So spending tree.

55
00:05:42,750 --> 00:05:44,890
Notice the root port.

56
00:05:44,920 --> 00:05:51,390
Gigabit 0 1 is in the listening state so we have listening.

57
00:05:51,390 --> 00:05:57,350
Then we have a learning and after a while it should go to forwarding.

58
00:05:57,660 --> 00:06:00,070
But that can take 30 seconds.

59
00:06:00,410 --> 00:06:01,800
So it's still learning.

60
00:06:02,040 --> 00:06:07,320
Now it's gone to forwarding and now Ping's will succeed.

61
00:06:07,320 --> 00:06:11,520
So ports have different states in a blocking state.

62
00:06:11,520 --> 00:06:13,740
Use a traffic is not forwarded.

63
00:06:13,830 --> 00:06:14,890
The switch doesn't.

64
00:06:14,970 --> 00:06:18,080
MAC addresses based on frame's received.

65
00:06:18,180 --> 00:06:22,510
This is a stable state for a port a listening and learning port.

66
00:06:22,590 --> 00:06:29,730
Do not forge frames either listing ports don't learn MAC addresses based on frames received.

67
00:06:29,730 --> 00:06:32,160
In other words they don't update the MAC address table.

68
00:06:32,340 --> 00:06:36,030
A learning port does update the MAC address table.

69
00:06:36,030 --> 00:06:43,880
This is a temporary state with transitionary state while the switches to learn the topology in affording

70
00:06:43,890 --> 00:06:45,610
state frames are forwarded.

71
00:06:45,900 --> 00:06:49,790
MAC addresses a Lent and this is a stable state.

72
00:06:49,800 --> 00:06:55,560
In other words this is not a transition restate the state will stay that way until there's a change

73
00:06:55,560 --> 00:06:56,950
in the topology.

74
00:06:56,960 --> 00:07:03,090
A disabled port doesn't receive frames doesn't Ford frames doesn't learn about MAC addresses on a port

75
00:07:03,650 --> 00:07:08,100
and this port will stay in that state until you enable the port.

76
00:07:08,100 --> 00:07:17,250
Now if we change that to rapid spending tree so spending tree mode rapid Peavey's t what we should notice

77
00:07:17,250 --> 00:07:20,760
is that convergence takes place a lot quicker.

78
00:07:20,780 --> 00:07:32,760
I'll only enable rapid spending tree on switch 1 2 and 3 show spending tree roll to show spanning tree

79
00:07:35,000 --> 00:07:41,020
shows us that the spanning tree mode now enables these rapid PV lines spending tree.

80
00:07:41,050 --> 00:07:51,510
We've got a pod cast using gigabit 0 1 so as you can see gigabit 0 1 is the root port.

81
00:07:51,550 --> 00:07:57,640
The switch can ping switch one I'll shut that port down.

82
00:08:00,010 --> 00:08:04,020
And when we do a ping again it instantly can ping.

83
00:08:04,060 --> 00:08:05,040
So which one.

84
00:08:05,080 --> 00:08:11,000
Even though we've just seen the interface go down in the output here because spanning tree converge

85
00:08:11,030 --> 00:08:15,070
is a lot quicker when using a rapid spanning tree.

86
00:08:15,130 --> 00:08:22,330
So the moral of the story is that in the real world you want to use a rapid previous t rather than previously.