Hi,
it seems like it's not possible to avoid having the elasticsearch_hosts in the server.conf set.
With this PR it's possible to leave it empty and as discussed in this Issue, the datanode hosts aren't necessarily coming from the elasticsearch_hosts-variable, at least as long as one is not using Basic Auth. This behaviour may cause problems in a clustered deployment, especially if the datanodes aren't on the same hosts as graylog-server. Only after some troubleshooting and removing the elasticsearch_hosts from the server.conf, we were able to start graylog-server and get to the preflight setup. Unfortunately, every time we run the playbooks using this ansible role, the elasticsearch_hosts gets set to either the default value or the supplied value. This causes graylog-server to not start properly and not listening on Port 9000
Hi,
it seems like it's not possible to avoid having the elasticsearch_hosts in the server.conf set.
With this PR it's possible to leave it empty and as discussed in this Issue, the datanode hosts aren't necessarily coming from the elasticsearch_hosts-variable, at least as long as one is not using Basic Auth. This behaviour may cause problems in a clustered deployment, especially if the datanodes aren't on the same hosts as graylog-server. Only after some troubleshooting and removing the elasticsearch_hosts from the server.conf, we were able to start graylog-server and get to the preflight setup. Unfortunately, every time we run the playbooks using this ansible role, the elasticsearch_hosts gets set to either the default value or the supplied value. This causes graylog-server to not start properly and not listening on Port 9000