Closed vichargrave closed 8 years ago
That's a very important setting indeed.
I take it by your answer that the discovery plugin does not take care of this and that this field should be exposed. If so, then I can add it and do a pull request.
On Aug 28, 2016, at 11:43 AM, Paulo Pires notifications@github.com wrote:
That's a very important setting indeed.
— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/pires/docker-elasticsearch/issues/18#issuecomment-242991663, or mute the thread https://github.com/notifications/unsubscribe-auth/AAP3vO8xiWxfYmwuzp-iee53IxLPAgP6ks5qkdbNgaJpZM4Ju-uF.
I was already expecting that.
OK I'll put the change in the docker-elasticsearch-kubernetes repo. I'm also going to add back the runs.sh script that was deleted from the 2.3.5 tag of that repo. Not sure why it was deleted.
Because it is inherited from this repo. Put it here, please.
ok will do. ignore my comment in that repo.
The solution here is specific to Elasticsearch, so discovery.zen.minimum_master_nodes
should be set in this repo.
Does the Kubernetes discovery mode of the fabric8 discovery plugin avoid split brain syndrome if you have multiple master nodes in your cluster? If not, then I'm wondering if the
discovery.zen.minimum_master_node
configuration field from a regular elasticsearch.yml file should be exposed in the Kubernetes Docker container elasticsearch.yml file.