Join GitHub today
GitHub is home to over 28 million developers working together to host and review code, manage projects, and build software together.Sign up
Plugin does not reattempt discovery after UnknownHostException in KubernetesClient #70
When using ElasticSearch 2.3.5 with the appropriate version of the Plugin, it seems that the API-DNS name is not discovered on the first attempt, which causes the join of the cluster to fail. Unfortunately the node does not reattempt joining the cluster, nor just shutdown the node in that case.
When manually restarting the pod it does recover the DNS gets resolved and the node joins the existing cluster.
This is the corresponding log.
Thanks for the plugin, it saves us a lot of time.
I am not sure, the plugin does return an empty list and does not throw the Exception.
I found out, that the Unicast provider in ElasticSearch, for AWS, Azure and GCE, do it in the same way. See the
So @pires is right, the plugin should not be responsible for retry.
I am sorry for the incovenience.