One of my customer has installed 2 MX65 with one as warm spare. A stack of MS250 is connected on MXes : 1 switch of the stack connected to active MX and the second switch connected to the warm spare.
When they simulate an hardware issue on the active MX, failover occurs, all hosts connected to the switches can continue to work properly.
When our cust wants to use manual switchover on dashboars, the active become spare and the spare becomes active, this is normal. But the users connected to the switch connected to the former active MX are stuck, unable to ping their gateway (MX). The only solution is physically restart the MX (former active, spare).
Tried to find out an explanation for my customer without success. Do you have any idea ?
do you have an additional failover link between the two mx devices? read this good blog: https://willette.works/mx-warm-spare/
is the switch management also lost? maybe the traffic is still send to the former active mx because the link was never down.
Check that they don't become active units at the same time. You mention this is not the case but just to double check... also check STP status in upstream links from the switches, and whether you have ALL vlans trunked between Firewalls or not.