Hello, Can you check ping between data centers? If delay of network may be long, you can incrase FailureDetectionTimeout (use org.apache.ignite.configuration.IgniteConfiguration#setFailureDetectionTimeout)
In additional you need make sure what communication ports are available in all nodes (by default org.apache.ignite.spi.communication.tcp.TcpCommunicationSpi#DFLT_PORT) . You can see all ports, which are will by used, in debug level logger. If nothing to help, please provide log files from all nodes and I could see details. Also I see you are not subscribed on users mail list. Can you please subscribe? You can do it from this page https://ignite.apache.org/community/resources.html Also I see you are not subscribed on users mail list. Can you please subscribe? You can do it from this page https://ignite.apache.org/community/resources.html visagan wrote > Hi, > > I have two nodes in data center A and two nodes in Data center B. I have > opened up the ignite default ports and i am able to talk to all nodes and > ports if i use "nc" command. > > But when i try to make the ignite cluster work, the four nodes join and > form a cluster and then they leave. > > I am able to see the total memory to increase from 24GB (when two nodes > are in the cluster) to 36 and then to 48GB and then the other two nodes > leave with an event NODE_FAILED. > But the nodes in the corresponding data center forms the cluster within > themselves and lives there happily. > When i turn on the cross data center, though it is able to discover the > nodes, it fails after joining with the other nodes. > Is there any possibly know reason for this ? -- View this message in context: http://apache-ignite-users.70518.x6.nabble.com/Cluster-Formation-between-nodes-in-different-data-centers-tp4838p4903.html Sent from the Apache Ignite Users mailing list archive at Nabble.com.