about Snapshots Questions

[wanghaiyongyw]

Cluster Full Storage Snapshot Recovery Questions:

1. While data is being written, if a full storage snapshot of the cluster is taken, how is the consistency of the restored data maintained?
2. Is it necessary to execute commands on every node for cluster snapshot backup?
3. Is the only way to perform a full restore of a cluster snapshot to load the backup snapshot file by restarting the service?
   Snapshot Recovery Issues Based on Collections:
4. When restoring a snapshot through the API, there is an option called "priority" with choices like "no_sync," "Replica," and "snapshot." What are the specific use cases for each?
5. Does this snapshot backup operation need to be executed on each node?
6. How should I plan and design the cluster to ensure data consistency with this type of backup?
7. What is the recommended backup method according to official recommendations? Can you provide detailed instructions specifically for snapshot backup?
8. Have you considered planning for network storage using Elasticsearch's snapshot method in the future?
9. Is it necessary for the number of replicas to match the total number of nodes in the cluster to address the data consistency issue during recovery? (This would mean that the data on each node in the cluster is the same)

Do we need to update the official documentation to version 1.4?