Scenario:
I have a switch (6248 without IP routing) which hangs off my core switch (6248 with IP routing) via cat5 cable, ports either side configured as trunk carrying vlan X.
Problem:
PowerConnect 6248 acting as Layer 3, created VLAN IP (directly connected, vlan X). Devices on the switch in vlan X work without issue, ping out, ping in, works. Devices on the "hang" switch, in vlan X seem to work, I can ping externally, receive pings from external sources, I see no issues with the service EXCEPT...
If I ping the VLAN IP(it's GW) from the device on the "hang" switch I get massive loss e.g.
64 bytes from x.x.x.x: icmp_seq=5 ttl=64 time=131.578 ms
64 bytes from x.x.x.x: icmp_seq=18 ttl=64 time=1.131 ms
64 bytes from x.x.x.x: icmp_seq=19 ttl=64 time=23.122 ms
64 bytes from x.x.x.x: icmp_seq=20 ttl=64 time=45.744 ms
I figure the switch receives the ping, goes to return it, thinks it's "directly connected" so doesn't try to pass it back over the trunk port, but that said **if I ping the device on the "hang" switch from an external source it works flawlessly**
Question:
Has anyone had this and corrected it, if so how and what stupid thing did I miss to created this headache?