link aggregation is not working properly on catalyst 9300 switch.

[n_lama]

i designed a network with link aggregation as shown, but there is link agregation problem i.e. link is getting down while removing one of the port on link aggregation.

waiting for support

[marce1000]

  - It depends on  what does link removal implies and how it is done ; always check logs on the switch where the link is removed and the remote partner , 

 M.

[n_lama]

I will check the detail log, but while i removed the one of the lan cable connected with port-channel on L2 9300 switch to active L3 9600 switch, then the connection is down for 5~6 seconds, and there is no down time while I removed the lan cable connected to standby L3 9600 switch.

and similarly if removed the lan cable with port-channel connected from L2 9300 switch to server then again there are no down time.

[Giuseppe Larosa]

Hello @n_lama ,

>> I will check the detail log, but while i removed the one of the lan cable connected with port-channel on L2 9300 switch to active L3 9600 switch, then the connection is down for 5~6 seconds, and there is no down time while I removed the lan cable connected to standby L3 9600 switch.

How are you doing your tests ? are you pinging  a server that is downstream the L2 switch or are you pinging the L2 switch management IP address ?

the 0 downtime on standby L3 switch can be expected if actually the packets are not going through it.

However, 5-6 seconds downtime is strange.

So it is better that you provide more details on your tests and your test traffic.

Hope to help

Giuseppe

[paul driver]

Hello
Can you elaborate on this please?

[n_lama]

@paul driver 

while i removed the one of the lan cable connected with port-channel on L2 9300 switch to active L3 9600 switch, then the connection is down for 5~6 seconds, and there is no down time while I removed the lan cable connected to standby L3 9600 switch.

and similarly if removed the lan cable with port-channel connected from L2 9300 switch to server then again there are no down time.

[MHM Cisco World]

use 
no lacp suspend-individual <<- if one link is remain in PO the PO is suspend this command will disable default behave
to be sure this is case here 

share 
show lacp internal<<- in both L3 SW
show port-channel summary <<- both side SW
show ehterchannel summary <<- both side SW

MHM