Spanning-Tree: The Dell Network Engineer who originally configured the stack disabled spanning tree for any ports connected to a Cisco Router or Sonicwall Firewall. I can't tell you exactly why that was done. What do you suggest?

% of Packet Loss: Pinging from a computer to the switch equals zero loss. Pinging from a computer to the default gateway router (both of which are connected to the Dell Switches) results in a 18-22% packet loss.

Subnet: The only subnet that I have tried is 172.18.91.0 as that is our primary local lan address.

Another Computer: Yes, same packet loss as above.

Received an invalid ip packet on: Those ip addresses are within our dhcp range, and would be coming from the HP Switch LAGs.

The document you sent deals with connecting Cisco Catalyst and Dell PowerConnects. We are not running any Catalyst switches. The Cisco I am connecting to is a Cisco 2911 Router which is running ip cef protocol.

VLANs: Vlan 1 is the only one now that handles client to client traffic.

Testing: No, I tried on three different switches. They are connected via ports Te1/0/43-46, and Te2/0/4

3-46 respectively.

It's been like this since initial setup, but this is the first chance I have had to dig into it. We got everything up and running with the PowerConnects/vmare/compellent san/dr4100/replication wise back in April, but then we started upgrading all of our bandwidth between our sites. I had need of port mirroring yesterday on the Dell switches which reminded me we still had to deal with this issue.

Based upon your firmware notes, I believe the next step would be to upgrade the firmware instead of testing another vlan. What is the latest firmware revision?

Ok, I was able to update the firmware this weekend. After that I was able to ping the default gateway with no issues, but then I wasn't able to ping the router or any server directly connected to it. So I went back to your next troubleshooting step.

"We should continue testing by taking two clients and plugging them into unused ports on switch 2 of the stack. Place the ports in access mode for something other than VLAN 1, and then  assign the test clients according IP/subnet. Then test pinging between the two devices."

I created a test vlan 99 and assigned it address 192.168.254.1. I have two laptops connected to ports TE 2/0/3 and TE 2/0/4 with ip addresses 192.168.254.101 and 102 respectively. I am able to ping from laptop to laptop with no drops, and each laptop can also ping the Dell switch without fail.

So the Dell switches still have trouble with handling the default vlan 1 traffic? That should not be the case.

That is our default subnet. Wouldn't changing it to a different vlan take down our network?

Ok, that's what I thought. Just making sure we were on the same page.

I may know the root of the issue, but I need clarification from Dell. Do the Dell switches automatically tag packets as the native vlan if no other vlan tagging is assigned? I ask because the HP switches are set to untagged as default out of the box, so all of our vlan 1 traffic is untagged traffic.

So here is my theory... When I connect my computer to TE1/0/3 which is on the native vlan of 1 and attempt to ping something connected to an HP switch, the ping fails due to the Dell Switch sending a tagged packet. Since the HP Switch is looking for untagged packets, it doesn't 'recognize' that traffic and therefore drops the packets.

Does that sound correct to you? If so, then I need to change our HP switches to tag all vlan 1 traffic.

What do you think?