One L3 DISTRO Dropping Packets, Second L3 DISTRO is Clean

[dewey89]

  I've got a strange delay issue on one of our L3 switches used as a DISTRO.  Distro 2 doesn't show these issues, but 1 does.

  Using a continuous ping from Distro1 out to devices in each VLAN I get one dropped packet after about 20-21 good ones.  Distro2 doesn't exhibit this.

  When I ping from the devices back to the Distro I don't see dropped packets, but I see latency in a regular pattern.  All pings would complete in 1ms except the one that takes 4-5ms.  See pattern below.

  I've searched for routing loops, but don't see any.  What else would cause this behavior?

!!!!!!!!!!!!!!!!!!!!!.!!!!!!!!!!!!!!!!!!!!!.!!!!!!!!!!!!!!!!!!!!!.!!!!!!!!!!!!!!!!!!!!!.!!!!!!!!!!!!!!!!!!!!!.!!!!!!!!!!!!!!!!!!!!!.!!!!!!!!!!!!!!!!!!!!!.!!!!!!!!!!!!!!!!!!!!!.!!!!!!!!!!!!!!!!!!!!!.!!!!!!!!!!!!!!!!!!!!!.!!!!!!!!!!!!!!!!!!!!!.!!!!!!!!!!!!!!!!!!!!!.!!!!!!!!!!!!!!!!!!!!!.!!!!!!!!!!!!!!!!!!!!!.!!!!!!!!!!!!!!!!!!!!!.!!!!!!!!!!!!!!!!!!!!!.!!!!!!!!!!!!!!!!!!!!!.!!!!!!!!!!!!!!!!!!!!!.!!!!!!!!!!!!!!!!!!!!!.!!!!!!!!!!!!!!!!!!!!!.!!!!!!!!!!!!!!!!!!!!!.!!!!!!!!!!!!!!!!!!!!!.!!!!!!!!!!!!!!!!!!!!!.!!!!!!!!!!!!!!!!!!!!!

[omz]

Hi

To me it looks like drops show a pattern .. ie. drop is always after certain replies. 

are there any port-channels? looks like port-channel issue to me .. 

[dewey89]

Yes, this is a standard collapsed core with two L3 switches connected to nine access switches and a TB out to the WAN.  There is a 2GIG port channel between L3 devices.  What port-channel issues cause this?  I really haven't heard of problems like this.