I have three new Cisco SG-300s operating in default settings. These are connected to two WatchGuard M300s in a HA configuration, with the WatchGuards acting as the router. These three switches connect to ten rack servers.
When I connect to the server's ILO NICs via one of the Cisco switches, 99% of packets are lost. I am able to browse each server's ILO GUI for around 30 seconds, after which, all network connectivity drops. This is the same whether I am connected via Remote Desktop into a server within the physical network, or whether I am using my laptop externally via VPN.
I have had extensive calls with WatchGuard, and after enabling debug logs across the cluster, I can see that each packet is allowed to transmit through the firewall to the switch and visa versa.
What confuses me is that if I connect to the WatchGuard cluster and use the inbuilt ping tool, 100% of packets are sent/received to the devices on the Cicso ILO switch.
I believe have ruled out the possibility of two new WatchGuards being at fault, and I cannot see a reason for each server's ILO port to be faulty, nor for all of the new patch cables to be faulty. For debugging, I have enabled DHCP, and each server ILO NIC picks up an IP address from the WatchGuard.
The three IP ranges are:
Cisco switch 1 (blue cabling) 10.0.1.X/24, connected to eth1 on the WatchGuard cluster (can ping fine 100% of the time).
Cisco switch 2 (red cabling) 10.0.2.X/24, connected to eth2 on the WatchGuard cluster (can ping fine 100% of the time).
Cisco switch 3 (green cabling - ILO) 10.1.0.X/24, connected to eth3 on the WatchGuard cluster (drop 99% of packets).
My question therefore is what within the Cisco switch or networking configuration could be causing this?