Jose, >As usual, my JDBC connection points to the local node (where I am activating >the cluster on), but does not connect to the cluster port (11211) only to >the default 10800 port.
11211 is a connector port, which serves a different purpose. By default JDBC connections are accepted on port 10800. >should I not be pointing to a general 'cluster' endpoint, rather than a specific node >address and port? If you use a thin JDBC driver, then you should specify one of the cluster nodes. So, all queries will be executed through the node you connect to. And connection won't break until you bring down this node. If you want a JDBC client to connect to a cluster without binding to some specific server node, you can use a client driver, which will start a client node locally and will execute all queries through it. Here is documentation on the client driver: https://apacheignite-sql.readme.io/docs/jdbc-client-driver Denis ср, 28 февр. 2018 г. в 18:12, joseheitor <j...@heitorprojects.com>: > Hi Denis, > > I have now launched the 3 nodes on different boxes, and listed their > addresses in each config. > > The nodes seem to be aware of each other, as I notice the topology updating > when I bring additional nodes online/offline. > > As usual, my JDBC connection points to the local node (where I am > activating > the cluster on), but does not connect to the cluster port (11211) only to > the default 10800 port. > > If I lose any or all of the (2) remote nodes, my JDBC connection continues > to work. But if I lose the local node - my database connection fails. > > I wonder if the problem lies in my JDBC connection... should I not be > pointing to a general 'cluster' endpoint, rather than a specific node > address and port? (It seems somewhat logical that my database connection > fails when the node that it is pointing to goes down...?) > > Jose > > > > -- > Sent from: http://apache-ignite-users.70518.x6.nabble.com/ >