Interface error gauging ?

[tedauction]

Hello, I have been asked to analyse all of our Cisco switches/routers (300) to determine if the interface overruns and port errors are enough to warrant purchasing new, more powerful devices.

Can someone please tell me give me an idea of what sort of statistics I should be looking for on a per interface basis i.e. how many dropped packets over what period.

Or alternatively how would I go about assessing this ?

Thank you kindly for any information.

[Dennis Mink · Accepted solution]

Typically, if you do a show interface x/y for instance errors and drop are cumulative values. this indicates only that a problem exists or has existed but not when. worst case it,. you would need to clear the counters for an interface an monitor of an extended period.

in my opninion you'd be better off using a network management tool (solarwinds NPM, Observium or something like that), this will give you error rates  (i.e. errors/drops against time) which is far more useful than looking at cumulative values.

[Mark Malone · Accepted solution]

Hi if your not familiar with calculating the error count through volume of traffic pushed

use Cisco Analyser tool from Cisco it will do it for you and also find other issues you may not be aware of , you can run the whole show tech from each router it will identify any problems  ----

http://www.cisco.com/c/en/us/support/web/tools-catalog.html

I would advise you to first clear all counters leave it for at least a week and then start

[tedauction]

Hello Mark, are you referring to the Cisco CLI Analyzer tool ?

I cannot see where it calculates error count - can you please advise where this feature is ?

Thank you kindly.

[Mark Malone]

sorry I got that wrong double checked this morning the old output interpreter could do it which moved to analyser  , now we use Cisco prime for that part under reporting which breaks it down ---Reports / Reports / Saved Report Templates / Interface Errors and Discards Report Details

you can calculate the old way , anything under 1% you dint need to worry about , with that many devices though you could be a while without automated software

GigabitEthernet2/29 is up, line protocol is up (connected)
  Hardware is C6k 1000Mb 802.3, address is 001c.58d3.cc7c (bia 001c.58d3.cc7c)
  Description: 
  MTU 1500 bytes, BW 1000000 Kbit, DLY 10 usec,
     reliability 255/255, txload 3/255, rxload 1/255
  Encapsulation ARPA, loopback not set
  Keepalive set (10 sec)
  Full-duplex, 1000Mb/s, media type is 10/100/1000BaseT
  input flow-control is off, output flow-control is off
  Clock mode is auto
  ARP type: ARPA, ARP Timeout 04:00:00
  Last input never, output 00:00:21, output hang never
  Last clearing of "show interface" counters 23w5d
  Input queue: 0/2000/0/0 (size/max/drops/flushes); Total output drops: 134505
  Queueing strategy: fifo
  Output queue: 0/40 (size/max)
  5 minute input rate 363000 bits/sec, 637 packets/sec
  5 minute output rate 15048000 bits/sec, 1262 packets/sec
     7888764354 packets input, 780031853491 bytes, 0 no buffer
     Received 49857 broadcasts (2 multicasts)
     0 runts, 0 giants, 0 throttles
     0 input errors, 0 CRC, 2 frame, 0 overrun, 0 ignored
     0 watchdog, 0 multicast, 0 pause input
     0 input packets with dribble condition detected
     15444718068 packets output, 23043085777544 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

divide total output drops by packets output will give you the percentage of drops over the length of the counter, 8.7% , o would need to work out why theres output drops , and how often , does it need qos , overruns usually indicate over utilization but output drops can be a number of causes