When B asks A (LEAVE-REQ) to leave the cluster, A sends a LEAVE-RSP to B, and B leaves when receiving this message (or a timeout ensues).
However, A seems to not stop sending the LEAVE-REQ to B when B rejoins (as B'). This requires TCP and fixed ports to be reproduced.
This goes on until UNICAST3.conn_expiry_timeout/conn_close_timeout has kicked in.
- causes
-
ISPN-13625 Scaling up 8.3 cluster with payload is slower compared to 8.2
- Closed
-
JDG-5031 Scaling up 8.3 cluster with payload is slower compared to 8.2
- Closed