09-01-2010 03:15 PM - edited 03-06-2019 12:46 PM
Greetings, we have high input discards daily
on our VPC's that go to 3750G access stacks. One 3750G is running
src-dst-ip and the other stack is src-mac, both are increasing the input discard counter on the Nexus side, but not on the 3750 side. I have 4948-10G and UCS VPC's that are totally clean on both sides, just these 3750G's reporting issues.
The duplex settings are auto neg at 1GB. What does it mean when one side gets RX input discards and the other side is clean?
Thanks!
Kevin Moline
Nexus7K-A# sh vpc statistics vpc 21
port-channel21 is up
vPC Status: vPC number: 21
Hardware: Port-Channel, address: 0024.f71d.3914 (bia 0024.f71d.3914)
Description: *** port channel to CISCOL2-02 stack ***
MTU 1500 bytes, BW 2000000 Kbit, DLY 10 usec,
reliability 255/255, txload 1/255, rxload 1/255
Encapsulation ARPA
Port mode is trunk
full-duplex, 1000 Mb/s
Input flow-control is off, output flow-control is off
Switchport monitor is off
Members in this channel: Eth10/5, Eth10/6
Last clearing of "show interface" counters never
30 seconds input rate 903672 bits/sec, 396 packets/sec
30 seconds output rate 3085632 bits/sec, 792 packets/sec
Load-Interval #2: 5 minute (300 seconds)
input rate 967.82 Kbps, 198 pps; output rate 3.00 Mbps, 693 pps
RX
2499364493 unicast packets 22547072 multicast packets 42035785 broadcast packets
2563947350 input packets 807335713881 bytes
0 jumbo packets 0 storm suppression packets
0 runts 0 giants 0 CRC 0 no buffer
0 input error 0 short frame 0 overrun 0 underrun 0 ignored
0 watchdog 0 bad etype drop 0 bad proto drop 0 if down drop
0 input with dribble 20163146 input discard
0 Rx pause
TX
3379827328 unicast packets 579669388 multicast packets 819525878 broadcast packets
4779022594 output packets 1823328519547 bytes
0 jumbo packets
0 output error 0 collision 0 deferred 0 late collision
0 lost carrier 0 no carrier 0 babble
0 Tx pause
1 interface resets
Nexus7K-A# sh vpc statistics vpc 22
port-channel22 is up
vPC Status: vPC number: 22
Hardware: Port-Channel, address: 0024.f71d.3910 (bia 0024.f71d.3910)
Description: *** port channel to L2 9-stack ***
MTU 1500 bytes, BW 4000000 Kbit, DLY 10 usec,
reliability 255/255, txload 1/255, rxload 1/255
Encapsulation ARPA
Port mode is trunk
full-duplex, 1000 Mb/s
Input flow-control is off, output flow-control is off
Switchport monitor is off
Members in this channel: Eth10/1, Eth10/2, Eth10/3, Eth10/4
Last clearing of "show interface" counters never
30 seconds input rate 18598536 bits/sec, 8613 packets/sec
30 seconds output rate 43301808 bits/sec, 15642 packets/sec
Load-Interval #2: 5 minute (300 seconds)
input rate 21.55 Mbps, 8.91 Kpps; output rate 54.20 Mbps, 16.53 Kpps
RX
127720532005 unicast packets 32962230 multicast packets 111702675 broadcast packets
127865196904 input packets 88136394079630 bytes
0 jumbo packets 0 storm suppression packets
0 runts 0 giants 0 CRC 0 no buffer
0 input error 0 short frame 0 overrun 0 underrun 0 ignored
0 watchdog 0 bad etype drop 0 bad proto drop 0 if down drop
0 input with dribble 13442373 input discard
0 Rx pause
TX
312987953559 unicast packets 568610668 multicast packets 749863821 broadcast packets
314306428048 output packets 239233492006027 bytes
0 jumbo packets
0 output error 0 collision 0 deferred 0 late collision
0 lost carrier 0 no carrier 0 babble
0 Tx pause
1 interface resets
Solved! Go to Solution.
09-02-2010 02:24 AM
Hello Kevin Moline,
Allowed vlans at the nexus side and the catalyst side are identical ?
There is a known issue that allowed vlan miss match causes "input discards" on nexus side.
Incoming traffic that has dot1q tag which is not allowed at nexus side is accounted in "input discard".
Regards,
Kimihito.
09-05-2010 10:37 PM
Hello Kevin Moline,
Are those "input discard" counters incrementing at a fixed rate, or you see bursty increments ?
From when it started incrementing ? Was there any change at that time ?
Also are you able to grab following outputs ?
# Nexus
- Multiple shots of commands below during the increments.
show interface ethernet
show interface ethernet
show hardware internal errors module
show hardware internal statistics pktflow dropped
show interface ethernet
show hardware internal statistics module
- Single shot of commands below.
show interface ethernet
show interface ethernet
show system internal ethpm info interface ethernet
show version
show port-channel summary
show interface trunk
show interface brief
show vpc consistency-parameters vpc 21
show vpc consistency-parameters vpc 22
# Catalyst
- Multiple shots of commands below during the increments.
show interfaces gigabitEthernet
show interfaces gigabitEthernet
show interfaces gigabitEthernet
- Single shot of commands below.
show interfaces gigabitEthernet
show interfaces gigabitEthernet
show version
show interfaces trunk
Thanks and regards,
Kimihito.
09-13-2010 11:09 PM
Hello Kevin Moline,
Could you have a chance to check the private message that I sent a couple of days ago ?
Regards,
Kimihito.
09-02-2010 02:24 AM
Hello Kevin Moline,
Allowed vlans at the nexus side and the catalyst side are identical ?
There is a known issue that allowed vlan miss match causes "input discards" on nexus side.
Incoming traffic that has dot1q tag which is not allowed at nexus side is accounted in "input discard".
Regards,
Kimihito.
09-02-2010 07:58 AM
Hello, thanks for the quick reply. Both ends are full trunk.
(2) member stack 3750G Po21 -> Nexus7ks po21 (VPC21) full trunk
(9) member stack 3750G Po22 -> Nexus7ks Po22 (VPC22 full trunk
09-05-2010 10:37 PM
Hello Kevin Moline,
Are those "input discard" counters incrementing at a fixed rate, or you see bursty increments ?
From when it started incrementing ? Was there any change at that time ?
Also are you able to grab following outputs ?
# Nexus
- Multiple shots of commands below during the increments.
show interface ethernet
show interface ethernet
show hardware internal errors module
show hardware internal statistics pktflow dropped
show interface ethernet
show hardware internal statistics module
- Single shot of commands below.
show interface ethernet
show interface ethernet
show system internal ethpm info interface ethernet
show version
show port-channel summary
show interface trunk
show interface brief
show vpc consistency-parameters vpc 21
show vpc consistency-parameters vpc 22
# Catalyst
- Multiple shots of commands below during the increments.
show interfaces gigabitEthernet
show interfaces gigabitEthernet
show interfaces gigabitEthernet
- Single shot of commands below.
show interfaces gigabitEthernet
show interfaces gigabitEthernet
show version
show interfaces trunk
Thanks and regards,
Kimihito.
09-07-2010 07:31 AM
Hello, thanks for the response. I will be posting this information today.
Kevin
09-08-2010 08:23 AM
I have the files that you requested, how do you want me to get them to you? There are several files...
09-08-2010 04:06 PM
Hello Kevin Moline,
I just sent a reply for the unicast message from you. Could you pls check your mail box ?
Regards,
Kimihito.
09-08-2010 04:32 PM
I'm havine the same issue on a Pair of Nexus, no VPC just a trunk link. My input discards are 100-200k per hour.
TAC had me make sure my vlan allow lists match but still having the same issue.
I'm using this card
sh module 10
Mod Ports Module-Type Model Status
--- ----- -------------------------------- ------------------ ------------
10 48 10/100/1000 Mbps Ethernet Module N7K-M148GT-11 ok
Mod Sw Hw
--- -------------- ------
10 4.2(3) 1.3
And running System version: 4.2(3)
Just curious what version of code are you running?
09-13-2010 11:09 PM
Hello Kevin Moline,
Could you have a chance to check the private message that I sent a couple of days ago ?
Regards,
Kimihito.
11-15-2010 08:08 AM
I have seen this issue before as well - even with vlan's matching up on both sides of the link. (3750G to Nexus 7k - vPC).
Was there ever a resolution or explanation?
Brad
11-15-2010 08:53 AM
Very interesting, I'm seeing the same problem myself. I have an etherchannel between two Nexus 7K's (not vpc). On one switch its reporting high input discards on the RX side. I've checked the etherchannel which is configured with the same allowed vlans.
Each N7K then connects to a 3750 stack. The 3750 terminates our dfark fibre for inter site communication.
Discover and save your favorite ideas. Come back to expert answers, step-by-step guides, recent topics, and more.
New here? Get started with these tips. How to use Community New member guide