High latency and dropouts on Network

how many ports are there in use on the Cisco switch?

It's a SG300-10SFP and all 10 ports are in use, including 2 pair of ports as port channels 1 & 2.

I don't see spanning tree being the issue. How many other switches are connected to this switch. What is at the end of the two port-channels? I'd check the make sure spanning tree is enabled on all switches. Make sure this core switch is the root switch. Then double check all the member ports in your port-channels. Make sure they are all up.

There are quite a few switches connected and all of them have spanning tree enabled (RSTP).

The two port channels each has 2x switchports allocated and used as link aggregation LACP to another network switch, where one of that serves as an uplink to the outside network and the other to a node room. All of 4x switchports were up.

How do you force a switch/switchport to be root? If the SG is the core switch in a star topology, should all switchports be root or just nominate 1 of the switchports?

Sorry I'm just trying to get my head around the spanning tree and it feels a bit like black magic :)

To reiterate the issue I'm experiencing, there has been a high latency and intermittent dropouts on the entire network and seems to coincide with a high CPU usage on the core switch (>45%). And the CPU usage seems to increase dramatically when STP is enabled. I'm not sure if my deduction of the STP causing network issue is correct?

Thanks!

If you have a star layer 2 topology and the SG is the hub, it should be the root. You can force it to be the root by lowering it's priority lower than the other switches. This will most likely be on a per vlan basis, so you will need to make sure it covers all of your vlan as the root. The root switch will have all it's ports be designated ports. Designated ports are downstream from the root switch, whereas root ports are upstream. In other words, the switches that are not the root, will have at least one root port back to the root. The other ports will either be blocked, or designated ports flowing downstream to other switches.

Thanks for the response!

Just to clarify, the core/root switch should have ALL ports as Designated.

And any edge switches, let's say the uplink port connected directly to a port on the root switch, should be set to Root on the uplink port?

And each of the ports have to be designated per VLAN basis?

Yes, spanning tree will sort it all out, that that is how the ports will be assigned by spanning tree based on the placement of the root switch.

So back to the latency and drop out issue, do you think the Spanning tree caused it? Or would it be because of the COU usage by spanning tree?

I can't see spanning tree causing a lot of cpu usage. Spanning tree is control plane more so than data plain traffic. In other words, bpdu's wouldn't cause a hit on network performance.

Would that mean the Spanning Tree has detected loop(s) hence the CPU usage? I'm trying to find out the cause of network latency and dropout and seems hitting brick wall.

No, a loop would most likely take your network down. 40% CPU, I'd think wouldn't cause a traffic spike. Switches switches use CEF to avoid using to much cpu for traffic forwarding. I'd need to know details of your topology to determine if you have a loop somewhere.  A loop doesn't come and go.