ok, thanks for the clarification. I'm running this series in my lab as well.
These switches do not follow the concept of access and trunk switches. You simply assign VLANs to ports, either untagged (would be like an access VLAN or native VLAN) or tagged (would be a trunk).
Let's assume you want to assign a VLAN to a port that should connect clients. you would simply add the port to the VLAN like this:
vlan 102
name "Client Devices"
untagged 1/2,1/5
exit
To assign this VLAN on an uplink you would use the tagged key word like this:
vlan 102
name "Client Devices"
tagged 1/8
exit
There is no need to assign the management VLAN to ports, except for the uplink. If you need to assign multiple tagged VLAN's on the uplink, you simply add those tagged to the same port as well. you can have one untagged VLAN and multiple tagged VLAN on a port.
A link aggregation (we call this a trunk and this makes it confusing) is very simple:
trunk 1/1,1/2 trk1 lacp
The command above would create a dynamic (LACP based) link aggregation.
you then assign VLAN's (tagged or untagged) to the trunk port, in this case, trk1. trk1 to trkx is predefined and the x is platform dependent.
Regarding STP, this is really dependent on your setup. In newer deployments, using VSF, you do not need spanning tree, as you will do not have loops. Some customers enable spanning tree to avoid loops during failure situations.
If you do not use VSF and you have multiple redundant uplinks, it would make sense to enable spanning tree. But there is no general answer to this question. It really depends on your environment and how you set it up.