07-05-2012 06:29 AM - edited 03-07-2019 07:36 AM
Hello,
I've got a recurring them developing across several deployments of Catalyst 3750X series switches, I'm seeing large amount of output drops accross multiple ports but there are is no QoS deployed on any of these switches. (standard FiFO).
As it happens the ports in question are a member of EtherChannel groups which are connected to ESX Hosts.
I've also seen this behaviour also unexplained on another 3750X at a different site, however it's a single port and connects to an Upstream firewall.
However in both cases neither switch has any QoS or non-standard configuration, literally a couple of VLAN's and some EtherChannel groups.
Here's some output from the case involving the ESX servers:
Port-channel6 is up, line protocol is up (connected)
Hardware is EtherChannel, address is 1cdf.0f82.1607 (bia 1cdf.0f82.1607)
Description: EC to xxxx
MTU 1500 bytes, BW 2000000 Kbit, DLY 10 usec,
reliability 255/255, txload 1/255, rxload 1/255
Encapsulation ARPA, loopback not set
Keepalive set (10 sec)
Full-duplex, 1000Mb/s, link type is auto, media type is unknown
input flow-control is off, output flow-control is unsupported
Members in this channel: Gi1/0/7 Gi2/0/7
ARP type: ARPA, ARP Timeout 04:00:00
Last input never, output 00:00:00, output hang never
Last clearing of "show interface" counters never
Input queue: 0/75/0/0 (size/max/drops/flushes); Total output drops: 101803
Queueing strategy: fifo
Output queue: 0/40 (size/max)
5 minute input rate 5220000 bits/sec, 463 packets/sec
5 minute output rate 734000 bits/sec, 321 packets/sec
21741137888 packets input, 26157897110996 bytes, 0 no buffer
Received 86503 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
29037782410 packets output, 28866202605938 bytes, 0 underruns
0 output errors, 0 collisions, 0 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
Port-channel7 is up, line protocol is up (connected)
Hardware is EtherChannel, address is 1cdf.0f82.1608 (bia 1cdf.0f82.1608)
Description: EC to xxxx
MTU 1500 bytes, BW 2000000 Kbit, DLY 10 usec,
reliability 255/255, txload 1/255, rxload 1/255
Encapsulation ARPA, loopback not set
Keepalive set (10 sec)
Full-duplex, 1000Mb/s, link type is auto, media type is unknown
input flow-control is off, output flow-control is unsupported
Members in this channel: Gi1/0/8 Gi2/0/8
ARP type: ARPA, ARP Timeout 04:00:00
Last input never, output 00:00:01, output hang never
Last clearing of "show interface" counters never
Input queue: 0/75/0/0 (size/max/drops/flushes); Total output drops: 78999
Queueing strategy: fifo
Output queue: 0/40 (size/max)
5 minute input rate 5310000 bits/sec, 494 packets/sec
5 minute output rate 818000 bits/sec, 361 packets/sec
30953679655 packets input, 34855698496736 bytes, 0 no buffer
Received 86894 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
48418978841 packets output, 52065337614843 bytes, 0 underruns
0 output errors, 0 collisions, 0 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
Has anyone seen this before or have any thoughts on it? Or does anyone know how I can attempt to identify what traffic is being dropped?
Thanks,
07-05-2012 07:03 AM
Which version of image you have on these switches ??
07-05-2012 07:04 AM
wHICH VERSION OF ioS IMAGE YOU HAVE ON THESE SWITCHES ???
07-05-2012 07:09 AM
Mixed, the one with ESX servers connected is running c3750e-universalk9-mz.122-53.SE2.bin.
The other is running c3750e-universalk9-mz.122-58.SE2.bin
Thanks,
07-05-2012 09:43 AM
Hello Duncan,
if we compare the output drops to the total output packets we see that the rate is very low
101803
over
29037782410
packets output,
means
3,505880668247627384848910712669e-6
and
78999
over
48418978841
1,6315709643406521093756166438521e-6
There is no real impact of these output drops on user traffic, the outport drops may be related to ASIC limitations on contiguous ports served by the same chip ( like Gi1/0/7 and Gi1/0/8) or by the fact that the port channel traffic has to be sent out a physical GE port (the load balancing algorythm may use much more a member link depending on traffic properties and so the traffic faces the single member link 1GE speed), but we can say that you are not facing real issues.
I mean if these are the error rates you don't need to investigate anymore.
Hope to help
Giuseppe
07-05-2012 09:48 AM
clear the interfaces and check the values in a 10 minutes range..
Alessio
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