2022-02-10T10:41:12,348][WARN ][o.e.c.c.ClusterFormationFailureHelper] [es-node-1] master not discovered or elected yet, an election requires at least 2 nodes with ids from [ti0MftEaQk2lV0VMglBfTA, RampKFimRgqlgb09m-ZapA, k0-c9tKzRIeKWgxBioADnA], have only discovered non-quorum [{es-node-1}{RampKFimRgqlgb09m-ZapA}{ok6IkBWXScOUofCKJKkMpw}{10.128.162.21}{10.128.162.21:9300}{cdfhilmrstw}]; discovery will continue using [172.28.162.22:9300, 172.28.162.23:9300] from hosts providers and [{es-node-1}{RampKFimRgqlgb09m-ZapA}{ok6IkBWXScOUofCKJKkMpw}{10.128.162.21}{10.128.162.21:9300}{cdfhilmrstw}] from last-known cluster state; node term 52, last-accepted version 12542592 in term 45
To get the simple things out of the way first, have you checked if the nodes can talk with each other and are not blocked by something like a firewall?
Your Elasticsearch.yml looks fine as it is, along as you have at least one other node also defined as master
@mverbeek until yesterday the cluster was fine, running on recently upgraded version 7.16.1 for at least one week. When we tried to enable Security and restarted it, then we run into this problem. Do you think that maybe the discovery.zen settings might be responsible somehow?
It's quite impolite to ping other people to draw them into a conversation in which they have not already engaged. Please don't do that. We're all just volunteers here, we unfortunately can't answer every question.
@mverbeek after replacing discovery.zen settings with discovery.seed_hosts and specifying there specific network interfaces (the nodes have multiple network interfaces), the problem seems to be corrected and the cluster after some time when to status GREEN.
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.