Multicast in 1.2 - vlan and subnet differences and partial completion
So there’s recently been an upgrade to all of the switches where I work and many computers seem to be in incorrect subnets and vlans now. Which makes very little difference in practice except when imaging. Which works. Just slowly.
Or, in the case of fog 1.2, I suspect causes some tasks to either full out fail or only partially complete. I set a load of tasks going last night to test this and it seems quite bizarre; some rooms have fully failed but they actually have some hosts that didn’t fail. The multicast log shows that some even finished first - how is this even possible with multicast?! The pc checkin numbers correspond to retransmissions to hosts that are different in subnet from the previous address in a lost, usually, but is this actually correct in assumption?
The point is, im interested in what is actually happening, so I can look in the right areas before asking the central ict department for help (I need to be able to ask them to change specific ports to a vlan etc). Does each subnet correspond to a separate broadcast, essentially allowing hosts to potentially finish the tasks before other hosts in another subnet have finished?
Also, it seems lots of hosts in queues end up sticking at waiting for slots, even when left overnight. Weird one really. Id set about 90 going on unicast, stacked up in a queue for 10, and now most are still in a queue waiting for nothing!