Hi Leo,To be sure, I have run

Alexander Pickar · ‎01-23-2015

Did anyone encounter a similar problem?

Our customer is migrating his switching infrastructure and during the migration period, he wanted to move his old 2950 switches from old Catalyst 6513 to a new stack of 2960-X. Metalic Gigabit Ethernet port to Gigabit Ethernet port. We have discovered, that the link wouldn't come up.

After playing a lot, we have discovered, that only 2960-X Model revision number A0 has this problem. There are 2960-X switches with revision number B0 in the stack and if you connect the old 2950 to these switches, there's no problem at all.

- config doesn't matter, because we just need the link to come up.

- straight or cross-over cable doesn't matter.

- MDIX auto turned on or off doesn't matter, we have tried to turn MDIX off and use a cross-over cable (like in those old days, you know?)

- When we set the link speed to 100, the link would come up (but who wants to have two switches connected via FastEthernet)

Here's the interface status and physical characteristics (believe me, there IS a switch connected to the GE interface):

2960-X_stack#sh int gi 1/0/25
GigabitEthernet1/0/25 is down, line protocol is down (notconnect)
Hardware is Gigabit Ethernet, address is 3462.88ca.0499 (bia 3462.88ca.0499)
Description: M.2.10.4
MTU 1500 bytes, BW 10000 Kbit/sec, DLY 1000 usec,
     reliability 255/255, txload 2/255, rxload 1/255
Encapsulation ARPA, loopback not set
Keepalive set (10 sec)
Auto-duplex, Auto-speed, media type is 10/100/1000BaseTX
input flow-control is off, output flow-control is unsupported
ARP type: ARPA, ARP Timeout 04:00:00
Last input never, output 00:02:13, output hang never
Last clearing of "show interface" counters 00:00:07
Input queue: 0/75/0/0 (size/max/drops/flushes); Total output drops: 0
Queueing strategy: fifo
Output queue: 0/40 (size/max)
5 minute input rate 0 bits/sec, 0 packets/sec
5 minute output rate 83000 bits/sec, 19 packets/sec
     0 packets input, 0 bytes, 0 no buffer
     Received 0 broadcasts (0 multicasts)
     0 runts, 0 giants, 0 throttles
     0 input errors, 0 CRC, 0 frame, 0 overrun, 0 ignored
     0 watchdog, 0 multicast, 0 pause input
     0 input packets with dribble condition detected
     0 packets output, 0 bytes, 0 underruns
     0 output errors, 0 collisions, 0 interface resets
     0 unknown protocol drops
     0 babbles, 0 late collision, 0 deferred
     0 lost carrier, 0 no carrier, 0 pause output
     0 output buffer failures, 0 output buffers swapped out
2960-X_stack#
2960-X_stack#
2960-X_stack#sh controllers ethernet-controller gi 1/0/25 phy

GigabitEthernet1/0/25 (gpn: 25, port-number: 25)
-----------------------------------------------------------

Registers and Values
------------------------------------------------------------------------
0000: 3100 Control Register                      : 0011 0001 0000 0000
0001: 7949 Control STATUS                        : 0111 1001 0100 1001
0002: 0381 Phy ID 1                              : 0000 0011 1000 0001
0003: 5CB2 Phy ID 2                              : 0101 1100 1011 0010
0004: 11E1 Auto-Negotiation Advertisement        : 0001 0001 1110 0001
0005: C9E1 Auto-Negotiation Link Partner         : 1100 1001 1110 0001
0006: 006F Auto-Negotiation Expansion Reg        : 0000 0000 0110 1111
0007: 2001 Next Page Transmit Register           : 0010 0000 0000 0001
0008: 4628 Link Partner Next page Registe        : 0100 0110 0010 1000
0009: 0600 1000BASE-T Control Register           : 0000 0110 0000 0000
000A: 0800 1000BASE-T Status Register            : 0000 1000 0000 0000
000F: 3000 Extended Status Register              : 0011 0000 0000 0000
0010: 8001 EMI Address Register                  : 1000 0000 0000 0001
0011: 0006 EMI Data Register                     : 0000 0000 0000 0110
0012: 8602 Phy Control Two Register              : 1000 0110 0000 0010
0013: 1051 Loopback Control Register             : 0001 0000 0101 0001
0014: 0000 Receive Error Counter                 : 0000 0000 0000 0000
0015: 000C MI Control Register                   : 0000 0000 0000 1100
0016: 0F08 Phy Config Register                   : 0000 1111 0000 1000
0017: 3048 Phy Control Register                  : 0011 0000 0100 1000
0018: 0000 Interrupt Mask Register               : 0000 0000 0000 0000
0019: 0140 Interrupt Status Register             : 0000 0001 0100 0000
001A: 0305 Phy Status Register                   : 0000 0011 0000 0101
001B: 1003 LED Control Register1                 : 0001 0000 0000 0011
001C: 00FF LED Control Register2                 : 0000 0000 1111 1111
001D: 1855 LED Control Register3                 : 0001 1000 0101 0101
001F: 03FC Diags Status Register                 : 0000 0011 1111 1100

Auto-MDIX : On [AdminState=1 Flags=0x00052248]
FC-FF-FF, 3F-0F, 01-03
2960-X_stack#
2960-X_stack#

I am wondering if anyone has encountered such a strange situation. If not, perhaps this post will serve as warning.

It would be nice to know, what other issues are hidden in the 2960-X model revision A0.

Alexander

Leo Laohoo · ‎01-23-2015

Last clearing of "show interface" counters 00:00:07

Uhhhh ... you just cleared the counters and then you posted the output to the command "sh interface".

Last input never, output 00:02:13, output hang never

According to this, there was something going from the switch to the downstream device.

Alexander Pickar · ‎01-26-2015

Hi Leo,

I suppose, you are referring to the non-zero "output rate". Well, this is because of the previous test, where I have reconfigured the interface speed to 100 Mbps. And because this is a 5 minutes statistics, you still see a positive, decreasing number there.

Let me once again summarize, when the link between these two switches come up:

When the 2960-X is model revision number B0
When the speed of the interface is manually set to 100 (Fast Ethernet)

Strange, isn't it?

Here's another output - counters cleared, but I have been waiting a while before checking the interface status:

a7-mdf-s1#sh int gi 1/0/25
GigabitEthernet1/0/25 is down, line protocol is down (notconnect)
Hardware is Gigabit Ethernet, address is 3462.88ca.0499 (bia 3462.88ca.0499)
Description: M.2.10.4
MTU 1500 bytes, BW 10000 Kbit/sec, DLY 1000 usec,
     reliability 255/255, txload 1/255, rxload 1/255
Encapsulation ARPA, loopback not set
Keepalive set (10 sec)
Auto-duplex, Auto-speed, media type is 10/100/1000BaseTX
input flow-control is off, output flow-control is unsupported
ARP type: ARPA, ARP Timeout 04:00:00
Last input 00:24:29, output 00:24:36, output hang never
Last clearing of "show interface" counters 00:23:38
Input queue: 0/75/0/0 (size/max/drops/flushes); Total output drops: 0
Queueing strategy: fifo
Output queue: 0/40 (size/max)
5 minute input rate 0 bits/sec, 0 packets/sec
5 minute output rate 0 bits/sec, 0 packets/sec
     0 packets input, 0 bytes, 0 no buffer
     Received 0 broadcasts (0 multicasts)
     0 runts, 0 giants, 0 throttles
     0 input errors, 0 CRC, 0 frame, 0 overrun, 0 ignored
     0 watchdog, 0 multicast, 0 pause input
     0 input packets with dribble condition detected
     0 packets output, 0 bytes, 0 underruns
     0 output errors, 0 collisions, 1 interface resets
     0 unknown protocol drops
     0 babbles, 0 late collision, 0 deferred
     0 lost carrier, 0 no carrier, 0 pause output
     0 output buffer failures, 0 output buffers swapped out
a7-mdf-s1#

Regards,

Alexander

Leo Laohoo · ‎01-26-2015

Alright Alexander. Let's try something else. But before we do, let me share you a unique "bug" with the 2960X: TDR.

WARNING: 2960X, and only the 2960X, somewhat support TDR. When I mean "somewhat" I mean your interface has to be 1000BaseTx. Unlike other switches, such as the 2960/G/S, 3560G/E/X, 3750G/E/X, 2960X (as of the latest IOS of 15.2(3)E) do NOT support TDR for 10/100BaseTx. (Please do not ask me why, you'll make me cry.)

So if this interface is a 1000BaseTx, do the following commands:

1. Command: test cable tdr interface Gi1/0/25;

2. Wait for approximately 5 to 7 seconds for the test to run;

3. Command: sh cable tdr interface Gi1/0/25; and

4. Post the output to #3.

Alexander Pickar · ‎01-26-2015

Hi Leo,

To be sure, I have run the test twice.

I don't think this to be a cable problem:

- When connecting the remote 2950 to a 2960-X, revision B0, the link comes up.

- Of course, other cables were tested prior to blame the switch.

- Also, the two switches are in the same server room, no structured cabling between these two devices.

Hey, but thanks for the command, didn't know, that it would be there.

Here are the outputs:

a7-mdf-s1#sh cable tdr interface Gi1/0/25
TDR test last run on: January 26 13:23:29

Interface Speed Local pair Pair length        Remote pair Pair status
--------- ----- ---------- ------------------ ----------- --------------------
Gi1/0/25 auto Pair A     N/A                N/A         Normal
                Pair B     N/A                N/A         Normal
                Pair C     N/A                N/A         Normal
                Pair D     N/A                N/A         Normal

a7-mdf-s1#sh cable tdr interface Gi1/0/25
TDR test last run on: January 26 13:24:44

Interface Speed Local pair Pair length        Remote pair Pair status
--------- ----- ---------- ------------------ ----------- --------------------
Gi1/0/25 auto Pair A     N/A                N/A         Normal
                Pair B     N/A                N/A         Normal
                Pair C     N/A                N/A         Normal
                Pair D     N/A                N/A         Normal
a7-mdf-s1#

Leo Laohoo · ‎01-26-2015

Thanks for the output.

I hate this restriction of the 2960X in regards to TDR only works on 1 Gbps. Output shows that the link is anything BUT 1 Gbps and more like 10 Mbps or 100 Mbps.

If the link is a point-to-point link and doesn't go to a distribution panel, do this:

1. Default a nearby configuration to the same switch and make sure STP portfast is turned OFF;

2. Plug the cable to the same switch.

When you plug the cable back to the same switch, you form a loop. With Portfast off, you shouldn't see anything bad happening. What you want to see is the link light. If the link doesn't go up, then you've got a cable problem.

It is a crude method, but there's no choice. I wanted to try this to eliminate a fault with cabling.

Alexander Pickar · ‎01-30-2015

Hi Leo,

I have asked the customer to connect two ports of the remote 2950 switch to the stack of 2960-X switches.

Gi 1/0/25 on 2960-x is on switch with revision A0.

Gi 2/0/25 on 2960-x is on switch with revision B0.

As expected, the port connected to 1/0/25 is down, 2/0/25 is up.

I swap the cables and once again, 1/0/25 is down, 2/0/25 is up.

If this was a cable problem, it would stick with the cable and not the switch-port.

Please see the attached file swap-cables-log_2015-01-30.txt.

(the port 2/0/25 goes to err-disable, but I don't care, I am happy, that line is up, protocol is whatever and switces see BPDUs).

My conclusion - there is an interoperability problem between Cisco 2950 and Cat 2960-X model revision number A0.

By the way, here's the TDR output:

a7-mdf-s1#sh cable tdr interface Gi1/0/25
TDR test last run on: January 30 12:13:32

Interface Speed Local pair Pair length        Remote pair Pair status
--------- ----- ---------- ------------------ ----------- --------------------
Gi1/0/25 auto Pair A     N/A                N/A         Normal
                Pair B     N/A                N/A         Normal
                Pair C     N/A                N/A         Normal
                Pair D     N/A                N/A         Normal
a7-mdf-s1#

a7-mdf-s1#sh cable tdr interface Gi2/0/25
TDR test last run on: January 30 12:14:07

Interface Speed Local pair Pair length        Remote pair Pair status
--------- ----- ---------- ------------------ ----------- --------------------
Gi2/0/25 1000M Pair A     24   +/- 10 meters Pair B      Normal
                Pair B     24   +/- 10 meters Pair A      Normal
                Pair C     24   +/- 10 meters Pair D      Normal
                Pair D     24   +/- 10 meters Pair C      Normal
a7-mdf-s1#

Alexander Pickar · ‎02-18-2015

The problem is resolved with the help of Cisco TAC.

The root cause is malfunctioning Energy-Efficient Ethernet on the interface.

After disabling EEE on the interface using command "no power efficient-ethernet" the link comes up.

2960-X Model revision number A0 fails to estabilish link