INDEX_NAME can not be imported as a dangling index, as index with same name already exists in cluster metadata

I have the following warnings in the log every day on a specific node:

(this repeats 285 times for the same 2 indexes)
[2017-01-19T00:53:30,171][WARN ][o.e.g.DanglingIndicesState] [LXS2jYx] [[card-views-and-searches-2016-12-05/-9uq5LO3RZ-zSAlzB_ojtg]] can not be imported as a dangling index, as index with same name already exists in cluster metadata
[2017-01-19T00:53:30,171][WARN ][o.e.g.DanglingIndicesState] [LXS2jYx] [[card-views-and-searches-2016-12-04/8wWRi8vsQFiQhIotXXLTWQ]] can not be imported as a dangling index, as index with same name already exists in cluster metadata

The logs seem to appear around the time our daily snapshot job runs:
[2017-01-19T00:53:29,818][INFO ][o.e.s.SnapshotShardsService] [dJor1ik] snapshot [s3-elasticsearch-snapshots:production-elasticsearch-backup-2017-01-19/-g85pwS2QRm3ALwY7gvS8w] is done

It seems it only happened to those 2 indexes.
card-views-and-searches-* is a index template

Environment details:

Production kubernetes cluster

Docker image
elasticsearch:5.0.0

Plugins:
RUN ./bin/elasticsearch-plugin install repository-s3
RUN ./bin/elasticsearch-plugin install io.fabric8:elasticsearch-cloud-kubernetes:5.0.0

3 elasticsearch containers with node affinity (1 per node) and with volume to the node fs

Anyone has any idea what caused this and how to fix it?

Thanks in advance.

This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.