I am not sure I can describe it any better than what is in the documentation you linked to. If you can explain what is not clear or confusing, I might be easier for me to help.
ok. I have some questions :
1.If I open an only one node in a cluster, then the node must be master node right?
2.data node only prepare for that if one day master node crash, then it would become master node,so usually it is only stand by and store data?
3.what does ingest node do?
These are all roles that a node can have. By default a node will have all roles, but you can, often when a cluster grows, start to have dedicated node types.
You must always have at least one node that is master eligible, but ideally 3.
Assigning a node the data role means that it can hold data. You must have at least one node that can hold data as well, otherwise you will have nothing to search. A data node can also be master eligible, but that has nothing to do with whether it holds data or not.
The ingest node role allows the node to process data before it is indexed.
I would recommend starting out with a 3 node cluster where all node have the default configuration (all roles) enabled. Once you outgrow this, you can start thinking about having nodes with dedicated roles. Just because you can have dedicated node types doesn't necessarily mean that you should.
ok !
so actually one node has all the function.the node type depending on how I set it.
so one master node could also have the data node attributes right?
Apache, Apache Lucene, Apache Hadoop, Hadoop, HDFS and the yellow elephant
logo are trademarks of the
Apache Software Foundation
in the United States and/or other countries.