Ether-channels can be a real pain if you configure them
in the wrong order, or trying to mix interfaces on different
blades, that happen to be a different rev hardware/software
wise.
Usually if you have configuration issues, the port(s) will
not come up in the bundle. They usually are in a stalled/wait
state.
Saying that, I was curious as to the IOS rev on these switches.
It looks like you're running older IOS code on switchs 1 & 3 and
switch 2 being the newer one, as per the configs you posted;
There could be a issue with ether-channel in the lower code
that may be causing the drops.
SW1#sh run
Building configuration...
Current configuration : 7628 bytes
!
version 12.1
SW2#sh run
Building configuration...
Current configuration : 7628 bytes
!
version 12.2
SW3#sh run
Building configuration...
Current configuration : 7628 bytes
!
version 12.1
There are a lot of issues over there years, if you
read the release notes for your IOS, you'd be amazed
at some of the crazy bugs that the IOS had pop up,
and been addressed in updates. One that was kind of
funny, was doing a "show run" command. That would
reboot a switch/router. Noone uses that command.
We've got a buttload of Ether-channels running in both
of our data centers, haven't seen the dropping issues.
Speeds range from 800MB, 8Gb, up to 20Gb channels,
in a mix of PAgP and LACP. The rootbridges for the vlan's
has been modified, so d1 is root for odd vlans, and d2
is root for even vlans. Down the road when we swap in
the nexus 7k's that's more than likely going to change, but
we'll cross that bridge when we come to it.
We have two camps here, one that nails up the ether-channels
hard, and a few white paper/best practice types that still
swear by the desirable modes.
one set examples, our data center here, has
dual 65xx distribution switches that dual attach to the
access layer switches. This is a smaller data center
as we've collapsed to a ton of HP C7000 blade centers,
so there's only 43 ether-channels left running, some
only have a single gig uplink running, but allows for adding
the second/fourth uplink on the fly.
Simple config for the Ether-channle between the two
looks like this;
!
!
!switch-D1
!
!
!
interface Port-channel102
description 20-gig-EtherChannel-to-switchd2
switchport
switchport trunk encapsulation dot1q
switchport mode trunk
mtu 9216
mls qos trust dscp
!
!
interface TenGigabitEthernet1/1
description switchd2-Te1/1
switchport
switchport trunk encapsulation dot1q
switchport mode trunk
mtu 9216
wrr-queue bandwidth percent 1 42 35 20 2 0 0
priority-queue queue-limit 10
wrr-queue queue-limit 5 45 20 10 10 0 0
wrr-queue threshold 1 100 100 100 100 100 100 100 100
wrr-queue threshold 2 100 100 100 100 100 100 100 100
no wrr-queue random-detect 1
no wrr-queue random-detect 2
no wrr-queue random-detect 3
wrr-queue cos-map 1 1 1
wrr-queue cos-map 2 1 0
wrr-queue cos-map 3 1 2
wrr-queue cos-map 4 1 3
wrr-queue cos-map 5 1 6 7
priority-queue cos-map 1 4 5
mls qos trust dscp
channel-group 102 mode desirable
!
interface TenGigabitEthernet1/2
description switchd2-Te1/2
switchport
switchport trunk encapsulation dot1q
switchport mode trunk
mtu 9216
wrr-queue bandwidth percent 1 42 35 20 2 0 0
priority-queue queue-limit 10
wrr-queue queue-limit 5 45 20 10 10 0 0
wrr-queue threshold 1 100 100 100 100 100 100 100 100
wrr-queue threshold 2 100 100 100 100 100 100 100 100
no wrr-queue random-detect 1
no wrr-queue random-detect 2
no wrr-queue random-detect 3
wrr-queue cos-map 1 1 1
wrr-queue cos-map 2 1 0
wrr-queue cos-map 3 1 2
wrr-queue cos-map 4 1 3
wrr-queue cos-map 5 1 6 7
priority-queue cos-map 1 4 5
mls qos trust dscp
channel-group 102 mode desirable
!
!
Ether-channel Summary
switchd1#sh etherchannel summary
Number of channel-groups in use: 43
Number of aggregators: 43
Group Port-channel Protocol Ports
------+-------------+-----------+------------------------
102 Po102(SU) PAgP Te1/1(P) Te1/2(P)
!
!
!switch-D2
!
!
!
!
interface Port-channel102
description 20-gig-EtherChannel-to-switchd1
switchport
switchport trunk encapsulation dot1q
switchport mode trunk
mtu 9216
mls qos trust dscp
!
!
interface TenGigabitEthernet1/1
description switchd1-Te1/1
switchport
switchport trunk encapsulation dot1q
switchport mode trunk
mtu 9216
wrr-queue bandwidth percent 1 42 35 20 2 0 0
priority-queue queue-limit 10
wrr-queue queue-limit 5 45 20 10 10 0 0
wrr-queue threshold 1 100 100 100 100 100 100 100 100
wrr-queue threshold 2 100 100 100 100 100 100 100 100
no wrr-queue random-detect 1
no wrr-queue random-detect 2
no wrr-queue random-detect 3
wrr-queue cos-map 1 1 1
wrr-queue cos-map 2 1 0
wrr-queue cos-map 3 1 2
wrr-queue cos-map 4 1 3
wrr-queue cos-map 5 1 6 7
priority-queue cos-map 1 4 5
mls qos trust dscp
channel-group 102 mode desirable
!
interface TenGigabitEthernet1/2
description switchd1-Te1/2
switchport
switchport trunk encapsulation dot1q
switchport mode trunk
mtu 9216
wrr-queue bandwidth percent 1 42 35 20 2 0 0
priority-queue queue-limit 10
wrr-queue queue-limit 5 45 20 10 10 0 0
wrr-queue threshold 1 100 100 100 100 100 100 100 100
wrr-queue threshold 2 100 100 100 100 100 100 100 100
no wrr-queue random-detect 1
no wrr-queue random-detect 2
no wrr-queue random-detect 3
wrr-queue cos-map 1 1 1
wrr-queue cos-map 2 1 0
wrr-queue cos-map 3 1 2
wrr-queue cos-map 4 1 3
wrr-queue cos-map 5 1 6 7
priority-queue cos-map 1 4 5
mls qos trust dscp
channel-group 102 mode desirable
!
!
Ether-channel Summary
switchd2#sh etherchannel summary
Number of channel-groups in use: 43
Number of aggregators: 43
Group Port-channel Protocol Ports
------+-------------+-----------+------------------------
102 Po102(SU) PAgP Te1/1(P) Te1/2(P)
The Port-channel interface
switchd1#sh int po102
Port-channel102 is up, line protocol is up (connected)
Hardware is EtherChannel, address is e8b7.48ed.0cf5 (bia e8b7.48ed.0cf4)
Description: 20-gig-EtherChannel-to-switchd2
MTU 9216 bytes, BW 20000000 Kbit, DLY 10 usec,
reliability 255/255, txload 1/255, rxload 9/255
Encapsulation ARPA, loopback not set
Keepalive set (10 sec)
Full-duplex, 10Gb/s, media type is unknown
input flow-control is on, output flow-control is off
Members in this channel: Te1/1 Te1/2
ARP type: ARPA, ARP Timeout 04:00:00
Last input never, output never, output hang never
Last clearing of "show interface" counters never
Input queue: 0/2000/0/0 (size/max/drops/flushes); Total output drops: 0
Queueing strategy: fifo
Output queue: 0/40 (size/max)
5 minute input rate 763133000 bits/sec, 89491 packets/sec
5 minute output rate 117069000 bits/sec, 36036 packets/sec
77036795986 packets input, 71787187179806 bytes, 0 no buffer
Received 528718397 broadcasts (469571224 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
47848107765 packets output, 34227653785063 bytes, 0 underruns
0 output errors, 0 collisions, 2 interface resets
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
I probably should have removed the QoS configurations, but
I wanted to show the difference between the physical interface
and the port channel interface.
Having all those Port-channels running VoIP and Video along
with various other mission critical apps, we can't afford
to have any uplink issues.
The only issue we have had, was with end-users plugging
the second network drop in the conference rooms into the
back of the VoIP phone, bridging both access layer switches
in that IDF. That resulted in a strange situation where
HSRP groups were hijacked.. but that's another story all
together.