Versions (relevant - OpenSearch/Dashboard/Server OS/Browser):
2.11
Describe the issue:
Hi,
Not an underlying issue although I require clarification around the management of shard counts on opensearch node servers.
We have three node servers. 3 primary shards and 1 replica shard. State management policy rolls over the indicie with the primary shard size is 30GB.
GET _cluster/health shows the below stats
“active_primary_shards”: 1407,
“active_shards”: 2748,
I understand that active_shards value is the count of the active shards and the replica shards although is this the value that I should focus on when taking into consideration the 1000 shard limit on node servers.
Further I recently made a change to the state management policy for the indices to close them after a length of time, expecting the shard count to decrease although that did not happen.
Are closed shards in an indicie counted towards the node shard count?
What is the preferred method to control this growth outside of deletion and adding another node server to the cluster.
I understand that I can increase the primary shard size rollover to 50GB although that will not help in the long term. Data needs to be retained for a year.
Thoughts?
Configuration:
Relevant Logs or Screenshots: