CBS350-24T-4x STACK LAG Configuration

[andywest]

I have two CBS350-24T-4x switches Stacked together using the 10G SFP+ ports in Native Mode.

I have created a LAG group comprising of 4 interfaces -  GE1 and GE2 from each of the units.
LACP is disabled 

When I connect the LAG group to a Cisco SG350X LAG group the network has difficulty processing packets and network applications stop responding.

I have switches in my network already connected using LAG groups and these are functioning fine. 
I suspect the LAG setup on the stacked switches as I have not utilized this method before. Maybe I have missed a step in the configuration.

I have assumed the interface allowing me to add interfaces from both units is acceptable.

The question is:

Am I allowed to utilize ports from both of the stacked units in a single LAG?

[marce1000]

 >...When I connect the LAG group to a Cisco SG350X LAG group the network has difficulty processing packets and network applications stop responding.

                - Anything in the logs when that happens ?

 M.

[Seb Rupik]

Hi there,

Can you confirm that the CBS switches are actually stacked together and not just connected in a ring?

To answer your question, having a LAG with member ports on different stack members is the preferred method of connection.

cheers,

Seb.

[andywest]

I removed the connections to bring the network back to a usable state.
I checked the switch stack management on the interface and noted that it was different from what I was seeing on the Internet how to guide.

I didn't mote any error in the logs.

It wasn't showing "Stack Mode" which sent me looking for a software update. Now that is installed it is showing Stack Mode as Native Stacking as I expected from the configuration I entered.
Stack Topology  = Chain

Stack Active = Unit 1

I can see both units linked  and can select the second unit so I'm confident that the units are stacked.

I have connected Unit 1 GE1 & 2 to the network and this is fine. I'm waiting till the end of the day to plug the others in and see what happens. It could be the Software level I was on.

[Flavio Miranda]

Yeah, it is allowed.

If you look at web interface manual for seting up LAB, it says:

Step 6. Choose the Unit of the switch from the Unit field which displays the stacking member for which LAG information is defined.

Which means,  LAG  is not limmited to one unit only.

Take a look on this material and try to identify if something can be done differently.

https://www.cisco.com/c/en/us/support/docs/smb/switches/Cisco-Business-Switching/kmgmt-2239-configuring-lag-settings-on-a-switch-through-cli.html 

https://www.cisco.com/c/en/us/support/docs/smb/switches/cisco-small-business-500-series-stackable-managed-switches/smb2860-link-aggregation-group-lag-management-and-settings-on-sx500.html?dtid=osscdc000283 

[andywest · Accepted solution]

Looks like I have found the problem, A patching error linking two ports not in the LAG group. 

The configuration on both switches was correct. The other switches in the network have interfaces incrementally across the top row then restarting on the second row. The new switches increment top then bottom then then next column.

The older SG switches have numbers on every row where as the CBS switches don't. Patching was carried out based on the SG configuration so instead of interface 2 being connected it was interface 3 on both switches in the stack.

I suspect this would cause spanning tree issues and hence the network slowness and dropouts.