I am setting up an Elasticsearch cluster in EC2 but it is failing with the crucial discovery stage. I am using packer to build the instances and this works. However when I start up a master I receive the message
"[2017-09-12T13:26:35,595][WARN ][o.e.d.z.ZenDiscovery ] [cF8JVs0] not enough master nodes discovered during pinging (found [[Candidate{node={cF8JVs0}{cF8JVs0iQx2BtY_ABnaaIg}{8Z1GmqjZQ0GJPwOSfSeCIQ}{10.9.32.188}{10.9.32.188:9300}{aws_availability_zone=eu-west-1b}, clusterStateVersion=-1}]], but needed [3]), pinging again"
even when three masters are up.
Our IAM policy is
{
"Statement": [
{
"Action": [
"ec2:DescribeInstances"
],
"Effect": "Allow",
"Resource": [
"*"
]
}
],
"Version": "2012-10-17"
}
and our EC2 settings are
discovery.ec2.node_cache_time: 120s
cloud.node.auto_attributes: true
discovery.zen.join_timeout: 90s
discovery.zen.ping_timeout: 30s
discovery.zen.hosts_provider: ec2
discovery:
ec2:
groups: [sg-67b5e51f]
host_type: private_ip
plugin.mandatory: discovery-ec2
network.host: ec2:privateIpv4
discovery.zen.minimum_master_nodes
The documentation appears to be a little out of date and I cannot understand what is going wrong.