Hi there,
We have a cluster on Windows where we're trying to launch Elasticsearch v7.7.1 . We are using cluster.initial_master_nodes
, setting it to the host-names (we think). We are not setting node.name
. We are being careful to only use bare hostnames, per https://www.elastic.co/guide/en/elasticsearch/reference/7.3/modules-discovery-bootstrap-cluster.html
We get the error message:
[2021-01-04T18:23:02,463][WARN ][o.e.c.c.ClusterFormationFailureHelper] [ANALYT1-PROD] master not discovered yet, this node has not previously joined a bootstrapped (v7+) cluster,
and this node must discover master-eligible nodes [Analyt1-Prod, Analyt2-Prod] to bootstrap a cluster:
have discovered [{ANALYT1-PROD}{pau7r42ARxaVPFvtH2HskQ}{9oWHmVkrSJkuVQpTASeCEw}{10.42.0.4}{10.42.0.4:8598}{dimr}, {ANALYT2-PROD}{PbXE1nPZRIKmekDYhPABgQ}{bwAUFvkqTSkt6oxs4-7hsQ}{10.42.0.5}{10.23.0.5:8598}{dimr}];
discovery will continue using [[fe80::b5f1:4242:c149:20e]:8598, 10.42.0.5:8598] from hosts providers and [{ANALYT1-PROD}{pau7r42ARxaVPFvtH2HskQ}{9oWHmVkrSJkuVQpTASeCEw}{10.42.0.4}{10.42.0.4:8598}{dimr}] from last-known cluster state;
node term 0, last-accepted version 61 in term 0
This is a little surprising, because we're used to DNS resolution not being case-sensitive.
However, we understand that this isn't DNS resolution. It just happens to be using host names because we aren't setting node.name .
Is the right workaround to use node.name to be sure that we're getting the matching case for both ? Or should Elasticsearch be more tolerant of case-sensitivity in the default-case where the node.name is a host-name?
Or are we facing another issue?