Hi, I'm a newbie to Ignite and am just trying to get a cluster setup. I have an Amazon EMR cluster with a bootstrap action that installs Ignite on each node and calls ignite/bin/ignite.sh on a sample configuration that uses the TcpDiscoveryMulticastIpFinder with default settings. I believe I have TCP/UDP access between slaves and master node on my EMR cluster. However when I start up another node manually, it reports only two nodes in my cluster (the two that are running on my master node) and doesn't seem to auto-discover the nodes running on the EMR slave nodes. I know there is an S3 IpFinder but I don't have permissions to create dedicated S3 buckets to anything right now so I can't use that. However given that the nodes in my EMR cluster have UDP access to each other, I would think that the default multicast should work. Am I missing a step somewhere? I've attached the sample config file I'm using. ignite-poc-config.xml <http://apache-ignite-users.70518.x6.nabble.com/file/n568/ignite-poc-config.xml>
-- View this message in context: http://apache-ignite-users.70518.x6.nabble.com/multicast-help-tp568.html Sent from the Apache Ignite Users mailing list archive at Nabble.com.
