Example of a snapshot error in Elasticsearch:
Global cluster state was stored, but at least one shard wasn't stored successfully...
Failed Shards: 2
From what I understand, the system in this case has made a correct snapshot on 8 shards?
If I want to have 10/10 shard successfully, do I have to remove the old snapshot (partial failure) and make a new one from the beginning?
Is it possible to make a snapshot "complementary" to these 2 shards with a partial failure? The system would only complete the error shards.
How does it look like?