I have a site where we tried to stagger trunks between multiple switches, for redundancy should one fail. We were hoping STP would deal with keeping things straight, however when we rebooted switch #2 in the below diagram, #3 went off permanently, as if it was only ever trying to look for a connection from its port50 and not failing over to port49.
On other sites we have switches that have dual trunks that FortiLink takes care of LACP automatically, like this:
Switch1-1 # show switch trunk
config switch trunk
edit "4FPTF20001720-0"
set mode lacp-active
set auto-isl 1
set members "port49" "port50"
next
end
But in the below picture example, here is the same output for Switch #1:
Switch2-1 # show switch trunk
config switch trunk
edit "8FFTF21004328-0"
set mode lacp-active
set auto-isl 1
set members "port48"
next
end
And here is the second switch's output in the below example:
Switch2-2 # show switch trunk
config switch trunk
edit "8FFTF21004165-0"
set mode lacp-active
set auto-isl 1
set members "port48"
next
edit "8FPTF20001565-0"
set mode lacp-active
set auto-isl 1
set members "port49"
next
edit "G101FTK19006567"
set auto-isl 1
set fortilink 1
set members "port47"
next
end
I'm just curious if this is not best practice, as something is definitely wrong-- a reboot of switch #2 takes #3 offline until the latter is power cycled.
This setup is below:
