VLAN X is 100% bandwidth utilization with lots of packets drop in my 6504

[itqan2003]

My monitoring system is reporting that VLAN 1 is 100% utilization with lots of packets drop in my 6504. That VLAN is used to route multicast traffic between 10 channel-ports each is 2G. It seems that by default the VLAN bandwidth is set to 1G, however I am passing more traffic than the VLAN default speed. I have "ip pim sparse-dense-mode" enabled inside this VLAN and all my 2.5Gbps multicast traffic is getting routed through it. Is it possible to increase the VLAN default speed to something higher than 1G? Or should I use "pim sparse-dense-mode" inside the physical ports instead of the VLAN? 

Thank you,

Alex

[Paul Chapman]

Hi Alex -

To make your monitoring application happy, just modify the bandwidth of your SVI.  It doesn't actually change the bandwidth, just calculations that are bandwidth dependent (like the monitoring system).

interface vlan1
 bandwidth 2500000

I'm not seeing drops in the port statistics you provided.  Are the drops occurring on the physical ports?

PSC

[itqan2003]

Hi Paul,

Thanks for the suggestion and I did this to cheat my way around the monitoring application. But if you look closely at the input queue there are 265211 discarded packets. 

Here's what the monitoring tools is showing: See the highlighted field in purple.    

Thanks again for your help,

Alex

[Ganesh Hariharan]

Hi Alex,

Usually Input Queue  can be seen when traffic arrving to SVI should be sent for CPU processing. If that traffic dropped by CPU - you will see the drops on VLAN input. Also if queue is full with these packets you will see flushes and no buffers.

Very good explanation along with some testing given on below link.

Input Queue Error

Hope it Helps..

-GI

[itqan2003]

Thanks for pointing this out to me. I don't have high CPU utilization but the memory is around 91%. I'll plan a maintenance window next month to run NetDR, along with wireshark capture to see which traffic is causing this.