Unassigned shards - problem (does not recover / assign)


I have a problem with the Elasticsearch cluster.

After restarting one of the nodes (resizing partitions in virtualization) this node started normally.

Some shards on the node work good.

A few shards, however, have a problem.

I have the status "Yellow" of few indexes (I have a number of replicas = 1).

However, the cluster does not want to fix these shards.

How can I fix it?

Try this Cluster Allocation Explain API, see https://www.elastic.co/guide/en/elasticsearch/reference/7.5/cluster-allocation-explain.html


"explanation" : "cannot relocate primary shard from a node with version [7.5.2] to a node with older version [7.5.1]"

7.5.2 is not compatible with 7.5.1 / 7.5.0?
It's normal?

you can run them in parallel, as querying and indexing works, but you should not do that permanently.

That allocation decider is currently just checking if the receiving node's version is smaller than the sender's version, and does not take major/minor versions into account.

Easiest thing would be to just get all nodes on the same version.

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