Peter,
first of all, interconnecting hp procurve switches to Cisco switches
may^Wwill introduce spanning-tree protocol problems.
Cisco-Default is one spanning tree inside each vlan.
HP default is "off" which may result in loops or one global spanning tree.
Solution here is to not build redundancy or
use MST on both, with identical mst configuration (name, versionnumber, assignment of vlans to mst instances).
Migration to it means downtime.
The other problem may be that some sorts of hp switches (2626, 2650 for example) are not able
to have the same MAC in different VLANs. (if you have an older sun server with multiple network interfaces,
you will have the systems-MAC on all network interfaces.)
On those switches, you see mac-address change from one port and vlan to the next,
all the logfile full.
And of course you may see outages.
This also happens when you have a cluster with ONE vrrp group on differnet physical interfaces:
on all that ports you see the same VRRP-MAC derived from the group-number.
(this seems to be the way checkpoint works.)
Also here you will get lots of messages about mac-address move/flapping.
So please look at your syslog or local logging on the switches.
Hope this helps,
Juergen.
