I am migrating my lab network from an old HP switch to an Aruba 1930 48 port switch. At the same time I have redesigned my network to take advantage of the routing capability between VLANs of the 1930. All has been going well - I moved across each group to their new VLAN and subnet (using several LACP trunks for NAS storage appliances) and had the DHCP relay point to the existing Win 2016 server (also providing DNS) on the old switch, and tested I could route between the subnets OK (I'll configure ACLs to provide finer control as the last piece of the puzzle).
However, when I finally brought the WIn 2016 server across to its own subnet/LACP trunk the switch went haywire, constantly rebooting. It just males no sense, but I figure I must have something wrong. Previously, I had to apply Storm Control and Loop Protection globally to prevent the CPU running at maximum - so maybe I have a loop somehow?
Does anyone have any clues? I have raised a support ticket but no response as yet.
@MrFixit Sounds like the last thing that was changed was adding your server, try removing it, and see if it stabilizes your network.
@GThiesen Absolutely, once I disconnect it the switch is stable again. I have support looking into the issue.
P.S. I have now disabled Storm Control, Loop Protection and STP both globally and individually by port/interface. The switch now runs at an idle, so i am not sure what else I must have changed to fix the earlier maxxed out CPU behaviour.
By way of an update, I've found that not using the Proxy ARP feature on my VLANs cures the crashing. I'm not sure why (my ignorance probably) but I don't think I need to use it anyway, so I guess it's a 'configuration issue' i.e. it was all my fault.
However, the DHCP Relay feature now doesn't work when I connect my Win 2016 server to the switch, so I'm not really any further forward.
Support are still investigating so I will have to see what they come up with.