I regularly replace network nodes using a life cycle schedule that helps ensure the best up time for patients and doctors and staff. And when I do, the new nodes are always different versions & models of Cisco hardware.
The new node takes over the management IP address of the old node, and NPM and NCM handle that just fine.
The problem comes when the new hardware is significantly different than the old hardware. Let's say a 24-port switch was replaced with a 48-port switch, or vise versa. Or perhaps a 100-Mb switch is replace with a Gig switch. This is where the problem shows up.
The old switch with Fa ports is replaced with a new one with Gi ports, and the old Fa ports remain in NPM. They're grayed out. And the port history for bandwidth utilization on the old Fa ports does not transfer to the new node's Gi ports.
The old grayed-out ports remain forever, apparently. The only way I've found to properly remove them is to delete the old node entirely and build a new one, and that's no way to maintain latency and bandwidth information across the years.
I'm facing this problem in greater scale today as I embark upon a project to replace 400 Cisco 2860S switches with a matching number of 3850's. I'll be deleting all the old nodes and rediscovering them as 3850's, unless there's some method I've not yet learned that will keep the port bandwidth and node latency stats with the new switches, while simultaneously removing grayed-out non-existent ports and moving their history to the new Gi ports on the 3850's.
Is this a thing?
Would this benefit your future network management needs?
All this needs is a great number of votes from Thwack people, and a positive attitude from Solarwinds!
