Below is the result when I run my curator action file
2017-06-19 16:05:34,455 INFO Preparing Action ID: 1, "create_index"
2017-06-19 16:05:34,465 INFO Trying Action ID: 1, "create_index": Create index for active diagnostic events
2017-06-19 16:05:34,469 INFO "<diagnostic-{now/d+1d}>" is using Elasticsearch date math.
2017-06-19 16:05:34,469 INFO Creating index "<diagnostic-{now/d+1d}>" with settings: {}
2017-06-19 16:06:04,492 INFO Action ID: 1, "create_index" completed.
30 seconds seems very slow and in many cases the operation times out.
Please understand that Curator only makes API calls to Elasticsearch. It's really an index-selection wrapper on simple API calls, nothing more. If Curator is taking 30 seconds to create an index, I would wager that it would take equally long if you were to create an index via curl or the Kibana Console/Dev Tools, because Curator effectively does the same thing with its simple create call. This indicates something in your cluster rather than Curator. The timeout is a giveaway, as the default timeout is 30 seconds.
I highly recommend investigating your Elasticsearch master node logs, and your cluster. How many shards do you have cluster-wide? How many per node? How many nodes do you have? Presuming you have an index template for the index being created, what are the default settings there? Elasticsearch is creating all primary and replica shards behind the scenes. As the cluster becomes more crowded, this amount of shard management can actually begin to take longer and longer. Without knowing more, my first guess is that this is what's happening in your cluster.
I believe you are correct. Investigating further I see corroborating evidence for what you say.
I should state that my testing of Curator was taking place on my local development environment (docker). I believe the single-node is struggling with resource allocation problems and is either choosing to not create resources or does so slowly.
I accept this is a result of my small environment.
Apache, Apache Lucene, Apache Hadoop, Hadoop, HDFS and the yellow elephant
logo are trademarks of the
Apache Software Foundation
in the United States and/or other countries.