Say that you have two nodes, one candidate node and one leader node. When you stop the candidate node and then
you restart it, the leader node locks_agent will reject the lock from the just restarted candidate node locks_agent because outdated (lower version). Is this the expected behaviour?
I was expecting the leader node locks_agent to remove all the associated locks from its state whenever it gets a down from a stopped locks_server process.
uwiger
commented
11 years ago
Say that you have two nodes, one candidate node and one leader node. When you stop the candidate node and then
you restart it, the leader node locks_agent will reject the lock from the just restarted candidate node locks_agent because outdated (lower version). Is this the expected behaviour? I was expecting the leader node locks_agent to remove all the associated locks from its state whenever it gets a down from a stopped locks_server process.