1
00:00:00,890 --> 00:00:09,200
Yes rather one which is this device and I'll ping router two from router one to see if ping succeed.

2
00:00:10,310 --> 00:00:14,300
So ping ten 112 pings succeed.

3
00:00:15,620 --> 00:00:22,910
But notice we are getting a message that there's a duplex mismatch between the router and the 2951 switch.

4
00:00:23,750 --> 00:00:27,020
So we've been told that there's a duplex problem here.

5
00:00:29,340 --> 00:00:35,970
And on the switch we're seeing that there's a duplex mismatch on fast Ethernet 203 to the router.

6
00:00:36,270 --> 00:00:43,710
So we've been told that there's a duplex problem here, but we're also told that there's a duplex issue

7
00:00:43,740 --> 00:00:45,840
on Fost Ethernet zero four.

8
00:00:46,830 --> 00:00:50,790
To the 2952 switch on Fast Ethernet zero four.

9
00:00:50,880 --> 00:00:52,980
So there's a problem here as well.

10
00:00:55,100 --> 00:01:01,370
Now you may not be physically connected to the console and hence you may not see these issues.

11
00:01:01,610 --> 00:01:08,960
That's a good reason to use a syslog server or a management application from a company such as SolarWinds.

12
00:01:09,140 --> 00:01:11,400
So let's ping rather to.

13
00:01:11,420 --> 00:01:13,220
Again, Ping succeeds.

14
00:01:14,640 --> 00:01:19,620
So pings aren't necessarily a good way to test for duplex mismatches.

15
00:01:19,980 --> 00:01:25,260
However, I'll do an extended ping to.

16
00:01:27,460 --> 00:01:28,210
That rather.

17
00:01:28,210 --> 00:01:29,560
So let's try that again.

18
00:01:30,700 --> 00:01:31,840
Ten one, one, two.

19
00:01:32,740 --> 00:01:39,400
I'll send a thousand pings and I'm going to make the data gram size a large size.

20
00:01:42,650 --> 00:01:48,590
So we want to send a lot of traffic onto the network and hopefully we'll encounter a problem.

21
00:01:49,040 --> 00:01:50,090
And there you go.

22
00:01:50,120 --> 00:01:54,020
We've got a message saying a late collision has occurred.

23
00:01:55,010 --> 00:01:57,230
So this is a problem you may encounter.

24
00:01:58,810 --> 00:02:04,960
One side may tell you that there are late collisions, as you can see over there.

25
00:02:05,080 --> 00:02:07,240
We're getting a number of late collisions.

26
00:02:07,990 --> 00:02:10,180
On the other side, we're not seeing that.

27
00:02:13,530 --> 00:02:16,980
So on the switch show interface if 0/3.

28
00:02:20,070 --> 00:02:23,190
We're not seeing a lot of elite collisions.

29
00:02:24,300 --> 00:02:31,350
The configuration of that interface is set to a speed of ten meg and duplex of full.

30
00:02:31,950 --> 00:02:38,460
But notice on this side show run interface f zero.

31
00:02:38,820 --> 00:02:42,600
The side has its speed set to 100.

32
00:02:43,380 --> 00:02:46,380
Let's confirm that show interface f is zero.

33
00:02:46,770 --> 00:02:49,560
Notice the duplex is half.

34
00:02:49,560 --> 00:02:53,850
Speed is set in this example to ten megabits per second.

35
00:02:53,970 --> 00:03:01,680
So the router because it can't negotiate properly with the other device, in this case, the 2950 switch

36
00:03:02,460 --> 00:03:08,970
has set its speed to a value of ten meg of duplex.

37
00:03:10,630 --> 00:03:18,130
So the point to remember is when there is a mismatch, the side that is set for half duplex.

38
00:03:19,100 --> 00:03:22,370
We'll see a late collisions on the connection.

39
00:03:22,370 --> 00:03:26,390
But this only occurs when you sending enough traffic.

40
00:03:27,680 --> 00:03:30,920
So extended ping doesn't show that problem.

41
00:03:30,950 --> 00:03:37,880
But when a large amount of traffic is sent on this half duplex connection, we're getting a late collisions.

42
00:03:42,220 --> 00:03:48,790
On the switch side show interface if 0/3.

43
00:03:50,550 --> 00:03:56,010
Notice the speed is full duplex, ten megabits per second, whereas on the router it's half duplex ten

44
00:03:56,010 --> 00:03:57,180
megabits per second.

45
00:03:59,670 --> 00:04:06,330
Even though I configured it to use 100 megabits per second, it's using a fail back position of the

46
00:04:06,330 --> 00:04:10,410
worst case scenario, which is ten meg of duplex.

47
00:04:10,440 --> 00:04:12,930
And in this case, we're having that problem.

48
00:04:13,200 --> 00:04:19,170
Notice this is 100 based TCF, whereas this side is 100 based TCS.

49
00:04:19,470 --> 00:04:26,250
So in this example, the router has gone for half duplex ten megabits per second, whereas the switch

50
00:04:26,250 --> 00:04:31,320
is using ten meg full duplex and hence we've got a duplex mismatch.

51
00:04:32,310 --> 00:04:37,590
So in summary, we get told by CDP that there is a duplex mismatch.

52
00:04:37,920 --> 00:04:43,620
The local interface is not using full duplex, whereas the other side is using full duplex.

53
00:04:44,040 --> 00:04:50,640
Pinging a device doesn't tell us that there's a problem, but when large packets are sent.

54
00:04:53,760 --> 00:04:56,820
User's main counter drops.

55
00:04:59,490 --> 00:05:03,060
Poor throughput and you may see late collisions.

56
00:05:03,990 --> 00:05:06,390
So let me saturate that link again.

57
00:05:09,110 --> 00:05:13,700
Send 1000 pings, large data, gram size.

58
00:05:17,620 --> 00:05:20,170
Show interface f0/3.

59
00:05:23,650 --> 00:05:28,870
Notice we are now seeing the late collisions taking place on the interface.

60
00:05:29,860 --> 00:05:31,570
So on the half duplex side.

61
00:05:32,750 --> 00:05:39,140
We see this output of late collisions on the interface show interface zero.

62
00:05:39,140 --> 00:05:41,690
Notice the number of late collisions is increased.

63
00:05:42,520 --> 00:05:49,840
On the full duplex side, we may get input errors such as CRC errors or runt frames.

64
00:05:50,140 --> 00:05:54,220
So in this example, we've got 28 input errors, 15 crosses.

65
00:05:54,400 --> 00:05:59,950
We aren't seeing any runs, but we are receiving duplex mismatch messages.