High Input Error & Overrun In Cisco ISR4431 Router

[Akhila Kumar Pradhan]

Hi,
We are getting high input errors and overrun in Cisco ISR4K router Gig port. we have replaced the cable and check with another Gig port but the result is same.
Could you please help on this.
Please find the attached show version and show interface details for your reference.

[Joseph W. Doherty]

This unit just have the base performance license?

If so, and if I'm reading your interface stats correctly, suspect you're hitting performance limits.

If you can enable evaluation license for high performance, try that.

[Akhila Kumar Pradhan]

Hi Joseph,

Thanks for the reply. Which license does it required for high performance, please specify the same. Generally Overrun issue occur when receiving capacity  of the interface was exceeded. But in my scenario its Gig Port and traffic hits some in Mbps. 

The issue occur only production hours only & in this port LAN is connected from a switch.

[Joseph W. Doherty]

I believe its a "high performance" license, but having not actually used an ISR 4431, I'm not 100% certain.

Out-of-the-box, the ISR is supposed to support 100 Mbps.  With additional licensing, up to 300 Mbps.

Yea, I see overruns in ingress, but the fact it's a gig port is what can create the issue.  I.e. the LAN can send bursts of traffic to the ISR faster than the ISR can handle.  Beside trying to upgrade the ISR performance (again, I believe you can enable a license in evaluation mode [time limited]), you could also try running that interface at 100 Mbps.  This should help avoid high speed bursts hitting the ISR.

[#5]

Hey Akhila, did you apply the platform hardware throughput license ? Regards Prasad

[lpassmore]

I have same issue on a 4431 router
Cisco IOS XE Software, Version 03.13.06.S - Extended Support Release
Cisco IOS Software, ISR Software (X86_64_LINUX_IOSD-UNIVERSALK9-M), Version 15.4(3)S6, RELEASE SOFTWARE (fc1)

#show interface

Hardware is ISR4431-X-4x1GE, address is 00a2.eefa.bac1 (bia 00a2.eefa.bac1)

Description: [Internet] ACET-FNN-803-087 Royal St (NextGen VS-VLL-002-593)

Internet address is 167.30.38.26/30

MTU 1500 bytes, BW 256000 Kbit/sec, DLY 10 usec,

reliability 255/255, txload 19/255, rxload 76/255

Encapsulation ARPA, loopback not set

Keepalive not supported

Full Duplex, 1000Mbps, link type is auto, media type is LX

output flow-control is off, input flow-control is off

ARP type: ARPA, ARP Timeout 04:00:00

Last input 00:00:11, output 00:04:09, output hang never

Last clearing of "show interface" counters 28w3d

Input queue: 0/375/0/0 (size/max/drops/flushes); Total output drops: 370

Queueing strategy: fifo

Output queue: 0/40 (size/max)

30 second input rate 76507000 bits/sec, 11041 packets/sec

30 second output rate 19666000 bits/sec, 6114 packets/sec

25914094732 packets input, 13278884432078 bytes, 0 no buffer

Received 1449007 broadcasts (0 IP multicasts)

0 runts, 0 giants, 0 throttles

15301971 input errors, 0 CRC, 0 frame, 15301971 overrun, 0 ignored

0 watchdog, 322267 multicast, 0 pause input

15831928709 packets output, 9969484229200 bytes, 0 underruns

0 output errors, 0 collisions, 3 interface resets

1742512 unknown protocol drops

0 babbles, 0 late collision, 0 deferred

3 lost carrier, 0 no carrier, 0 pause output

0 output buffer failures, 0 output buffers swapped out

We have 2x1Gb outward-facing interfaces, both of which are 100Mb managed services so we shouldn't be getting more than 200Mb (OK, maybe 400Mb if we are sending and receiving at full capacity). The throughput of the device should be 500Mb as per:

#show platform hardware throughput level
The current throughput level is 500000 kb/s

We do have a third (toward our network) interface running at a full 1Gbps.  Does the figures for this interface count in the throughput?  I wouldn't have thought so since it is essentially the same traffic as would be found on the other two outward-facing interfaces.

And, the errors are only occurring on one interface, never on the second.  If this was a performance licence issue, I would expect drops on both interfaces I think.

It also doesn't appear to be load related as I have looked at the utilisation graphs and the drops can occur at any time, not necessarily when the interface is loaded.

I have  never seen the CPU go much above about 30% but we are using outgoing NAT on both outward-facing interfaces.

Are there any other areas to explore?

Thanks