Unassigned shards because of different elasticsearch version

Elastic Stack Elasticsearch
1

Hi All,

Recently I have upgraded the one data node , Node got successfully upgraded but there are some unassigned shards which are not being replicated to other nodes because of below error:

"explanation" : "cannot allocate replica shard to a node with version [7.13.4] since this is older than the primary version [7.16.3]"

Cluster is in yellow state.

Can anybody help me in this ?

2

Is there any way to reassign those replica shards ?

3

Yes, you must complete the upgrade. From the docs:

Once you have started to upgrade your cluster to version 7.16.3 you must complete the upgrade.

4

As of now there is no way out to assign those replica shards?

5

Yes, just carry on with the upgrade.

6

After upgrading the few nodes, I can see logstash service got failed inside the machines. And we are unable to restart the service.

Please refer below logs:
[2022-03-17T07:44:42,432][ERROR][logstash.outputs.Elasticsearch][main][db1af4057ae53eec4aebe1d43e3d39bc64c27408ff34e47b0e3370f6927d375f] Attempted to send a bulk request but Elasticsearch appears to be unreachable or down {:message=>"Elasticsearch Unreachable: [https://logstash_writer:xxxxxx@elasticsearch.mgmt.aws.uk.tsb:9200/][Manticore::SocketException] Connection reset by peer (Write failed)", :exception=>LogStash::Outputs::Elasticsearch::HttpClient::Pool::HostUnreachableError, :will_retry_in_seconds=>2}

Please help on this.