I've moved a one pool member into graceful disabled state for maintenance and since then most of the clients have been switched over to other member but there are still three client IPs hanging on the old server. I suppose these clients stay up all nights long, so the sessions don't really die. Maybe an open Horizon Client refreshes connections at shorter intervals than LB releases persistence (stickiness) for an IP.
I've set the persistence timeout to 1800 seconds by purpose as I found out, that if a client was switched between gateways he had to login again, but if I kept him mapped to a same UAG during the day then a single login at the morning was usually enough. One could had half an hour break after which the connections still worked instantly; think about closing a lid of a laptop, having a lunch and then coming back within 30 mins.
But now this timeout value is backfiring me so is there a way to terminate sessions or clear persistence tables from the Edge CLI or API.
I'm using CLI command "vdi-nsx-edge1> get load-balancer <GUID> persistence-tables" to monitor these. Three IPs have been hanging on a one "going to maintenance" server for days now.
