High Packet Drop and connection failure

These types of problems are always like going down the rabbit hole, especially over a medium like this. It does look like we are starting to narrow it down.

 

 

The LTM1 <> LTM2 ping loss is worth looking into, but not yet. I would be more concerned about losing pins from LTM1/LTM2 to any node under the 2960. I want you to find a node that is actually plugged into 2960-1. This is important. We don't want that traffic going across the PAgP link. I then want you to perform two ping tests. One with a 1000 and one with 10000 pings issued from LTM to the internal/inside node that is plugged into 2960-1 physically. Please report back.

 

 

If you are getting TCP RSTs from the client, then you need to dig further into this. This is the clue we have been looking for. Please verify the RST is from the client first. Make sure there isn't one from the server beforehand and also verify what communication took place right before the RST. Sample a couple of RSTs to see if it is the same call causing this. Focus in using the tcp.port or udp.port filter in wireshark.

 

 

The client is probably sending TCP RSTs because you're not doing anything and you ending the session (QUIT command). Sounds like normal behavior to me.

 

 

BTW, what version are you running again? 10.2?