Single Machine Cluster Setup - Are my elasticsearch 5.0.0 configurations proper?


(Chandra Kanth Nadikuda) #1

I am using elasticsearch 5.0.0 version. And chose to go with single machine cluster setup so that I can scale it by adding more machines later when load increases.. but I get an error - after elasticsearch runs for a few days..

elasticsearch.yml: only one machine with IP : 10.0.1.159, 16GB RAM, ubuntu 14.04
cluster.name: elasticsearch
node.name: es-01
bootstrap.memory_lock: true
network.host: 10.0.1.159
discovery.zen.ping.unicast.hosts: ["10.0.1.159"]

I have set environmental variable: export ES_JAVA_OPTS="-Xms8g -Xmx8g"

sysctl -w vm.max_map_count=262144
and verified mmap is unlimited


My setup runs fine .. until after elasticsearch running for a few days.. I get this error from Java Transport client

NoNodeAvailableException[None of the configured nodes are available: [{#transport#-1}{08UFeRP9RmCH4jMoV0RCNw}{10.0.1.159}{10.0.1.159:9300}]]
at org.elasticsearch.client.transport.TransportClientNodesService.ensureNodesAreAvailable(TransportClientNodesService.java:314)

-- I should not be getting this error .. Am I doing something wrong with configuration
I doubt if I have to write the same machines' ip address in unicast hosts property...

Please help


(Igor Motov) #2

What do you see in the elasticsearch log file? Does elasticsearch instance still run at this point? How do you recover from this issue?


(Chandra Kanth Nadikuda) #3

At that point , i see that elastic search uses much of ram.. and the system starts to hang.. i restart elastic search service to get it working..


(Igor Motov) #4

I was asking about elasticsearch log files (not your application log files). How did you determine that elasticsearch uses too much of ram?


(system) #5

This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.