Hi, this is called "partition awareness" [1] - thin client establishes connections to all known nodes to be able to send requests directly to the primary node for the given entry. Multiple connections are also useful for load balancing and improved reliability.
You can disable this behavior with ClientConfiguration#setPartitionAwareness(false) [2] [1] https://ignite.apache.org/docs/latest/thin-clients/java-thin-client#partition-awareness [2] https://ignite.apache.org/releases/latest/javadoc/org/apache/ignite/configuration/ClientConfiguration.html#setPartitionAwarenessEnabled-boolean- On Tue, Sep 6, 2022 at 2:09 AM Gregory Sylvain <[email protected]> wrote: > > > Hi, > > > > I am running an Ignite 2.12.0 Service Grid with 5 Server Nodes and 32 Thin > Client nodes in a private cloud of RHEL 7.9 VMs. (no docker images). > > > > Doing a sqlline CLI query for connected clients, I get the expected 32 > connected clients. However, if I execute netstat (or ss) on all > ServerNodes, I see an ESTABLISHed connection from each client to each > ServerNode on port 10800? These connections seem to be maintained (e.g. > they do not timeout after 2 hours). > > > > I am using Spring XML to configure both client and server nodes. The > server nodes are also utilizing systemd for reliability. > > > > Any idea what is going on? > > > > Thanks in advance. > > Greg > > >
