Keywords: Elasticsearch - Virtual Machines - Technical issue - Other
I'm using VMWare Workstation 14 Pro.
I downloaded the Bitnami Elasticsearch 6.2.4 OVA with Debian 9 installed.
The first VM created from OVA came up fine and was running as expected.
I wanted to create a second node for the ES cluster so I created another VM from the same OVA.
The second OVA would not join the ES cluster because it was (apparently) using the same GUID as the first instance.
After much googling I found that I had to delete the bitnami/elasticsearch/data/nodes/_state directory and restart the node.
According to ES documentation the node GUID is generated at first startup.
It would appear the Bitnami OVA has a "stale" data directory from when it was created or configured.
Once I deleted the _state directory and restarted the node it generate a new unique GUID and joined the cluster no problem.
This is either a bug in the OVA creation or should be documented.
Hopefully this helps someone else.