LACP is for link aggregation.
2 ports on switch A and 2 ports on switch B.
2GB link between them with LACP.
If not using LACP or static team 2 ports connected between switches would create LOOP. This is what RSTP is for. To prevent loops and lear about topology changes:
Right. Got that part.
1. So you can`t have just 2 cables between switches and expect to have 2 GB if LACP or static teaming is not used
2. 2 cables between switches with no RSTP (STP) and no lacp or static team will result in a loop
3. 2 cables between switches with RSTP (STP) and NO lacp or static teaming will result in only 1 link up and second blocked to prevent loop
Got that part too. With #3, you get loop protection, but no additional usable bandwidth between the switches, and RSTP while faster than STP is still pretty slow.
4. 2 cables between switches WITH RSTP and WITH team will result in 2GB link and failover.
This is the statement that confuses me. Although maybe I'm reading more into that than is there. If running RSTP on top of LACP, I would think that the STP is going to block one port so you are NOT going to get 2GB of throughput. If it's not blocking one port, than what is it doing, or to put it another way, why should it be there?
As I said in the first post, in general terms, I understand what (R)STP does, and what LACP does, what I am lacking on is how or if they work together on a two connection link between two switches, and should I be using one vs the other vs both.
I know I can't just plug in two cables between the two switches as that would create a loop. As I understand it (R)STP will prevent a loop by blocking one of the links. In the event of a failure of one link, the (R)STP will simply use only the one working link. LACP will aggregate the two links functionally into one link of higher bandwidth which will prevent a loop and if one fails, the aggregated link will simply have less bandwidth. Further, that the link failure detection in LACP is far faster than RSTP or even more so with STP..
For my purposes, my primary objective of this is to have redundancy in the event of a failure of one link between the switches. I don't really need the additional bandwidth (although It can't hurt), but I do want a failure detected quickly so traffic is impacted for as short of amount of time as practical.