I have a support case open for this, but thought I'd post something here while waiting to hear back. We spun up a new data node in an existing cluster. The rest of the facts:
- The node joins the cluster
- The master reports (on the cluster pulse marvel dashboard) that it is relocating shards to it from one of the existing nodes.
- No shards move to the new node. There is no information about why the shards aren't relocating on either the data node where the shards are trying to move from, the currently elected master, or the new data node.
- I can share a folder on the data volume of the new node and write a file to that share from the data node where shards are trying to relocate from.
- There is enough disk space on the volume (3 TB free)
- elastic did create the F:\Elasticsearch\data\clustername\nodes\0 directory on the data volume, and in that directory is a 0 KB node.lock file.
The .yml is the same as on the other nodes in the cluster (except for the node name).
The version of ES and java are the same.
OS is the same
How can I get more detail about why data isn't actually relocating?