Hello - We'er considering moving our current EC2 Based Stack to Docker. High level topography
consists of 5 EC2 Instances with all of them acting as Master/Data Nodes. We have two More Coordinating nodes that direct traffic to the 5 master/data nodes. The Two coordinating nodes are fronted by an ELB. We do get approximately 1 million documents per day which this setup handles with ease.
A few questions here. Any insights would prove useful in us progressing:
Does it make sense to setup the master/data nodes as DAEMON (thus ensuring we get only 1 task per Instance mimicing our current 1 Node per instance in the EC2 setup?).
Will this setup support automatic rollover in case of EC2 AMI Updates (as in at least 1 task/node is remaining alive when others are down during AMI upgrade ?)
If you're installing from a DEB/RPM then you might as well use the included service/daemon.
If you have multiple nodes down at one time, then your cluster may be in a yellow or red state, that will impact the availability of the services it provides.
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.