Having 2 allocation IDs in in_sync_allocations when there are no replicas

Hi

I am facing this issue where I had an index with an initial replication of 1. However, one year ago, due to disk storage issues I reduced the replication to 0.

One year later, I now run this operation: GET /_cluster/state?filter_path=metadata.indices.my_index.in_sync_allocations.*,routing_table.indices.my_index.* and I realised that there are 2 allocation IDs in in_sync_allocations.0 even though there are 0 replicas. Tried to find documentation of this error but to no avail. Does anybody know what is happening here?

That’s ok, it just means you haven’t written to this shard since dropping the replica, so if the current primary fails then ES can still safely recover from either of the shard copies that existed back when you set number_of_replicas: 0 if they should still exist.

That’s great news! Thanks for the swift response! Cheers

1 Like