When I was exploring link down and controller restart, there was a case where current_path ended up using the same path as failover_path, but with a different svlan.
By design, failover_path is supposed to be disjoint with current_path, so this shouldn't happen. I only noticed this issue when I saw the EVC output that I collected during a test (after I no longer had the scenario up and running), so I don't know if it was just transitory and how the flows actually were at that time. Either way, I'll register this bug here for future ref and to be eventually confirmed or not, it's worth trying to explore more link down/up with controller restarts with EVCs that have failover:
When I was exploring link down and controller restart, there was a case where
current_path
ended up using the same path asfailover_path
, but with a different svlan.By design,
failover_path
is supposed to be disjoint withcurrent_path
, so this shouldn't happen. I only noticed this issue when I saw the EVC output that I collected during a test (after I no longer had the scenario up and running), so I don't know if it was just transitory and how the flows actually were at that time. Either way, I'll register this bug here for future ref and to be eventually confirmed or not, it's worth trying to explore more link down/up with controller restarts with EVCs that have failover: