strange trunk behavior

[JoshIT]

hi,

I have a Proliant 2626 (switch B) which is trunked to another 2626 (switch A). Switch A is then trunked to a Dell 5324 switch. A 100 Mbps port on A (port 10) is used to form the trunk with a gigabit port on B (port 26). Not optimal, but this is how it has been always and everything was working all right. The dell switch is a Gigabit switch, any port of which can do Gigabit. All port on these three are configured for auto neg.

So what happened was a firmware update was performed on the Dell. Following that, PC-1 connected to switch B began to show request timed out in pings to PC-2 connected to the Dell (or some other switch trunked to the Dell). The ping would go through, but would time out every now and then.

The solution was, moved the trunk (between A and B) on A to a gigabit port on A, port 25. Any non-gigabit port on A when created as a trunk with B would exhibit the same ping timed out behavior between PC-1 and PC-2. Nothing else has changed on any of the switches other than the firmware update on the Dell.

Can anyone offer any insight on this strangeness? Switch A also happens to be a stack master for the HP switches.Switch B has the latest firmware on it, but not A.

thanks!