How to migrate from version 7.12 to version 7.16.2?

[wajika]

7.12 cluster

7.16

I made a snapshot on the 7.12 cluster, shared the snapshot using NFS, then created a new 7.16 cluster, mounted the nfs path, and registered the warehouse with the same name on kibana. After the registration, I browsed the snapshot UI and found Didn't find any existing snapshots, is this the wrong way to do it?
I see that Elasticsearch cloud service has such an operation case, I don't know if it is also suitable for private deployment environment.

Also I have a question, will snapshots keep xpack authorization records?

[warkolm (Mark Walkom)]

Did you mount the repo on the newer cluster as read-only?

Restore a snapshot | Elasticsearch Guide [7.16] | Elastic

[wajika]

Current result

cat /proc/mounts | grep nfs
192.168.10.141:/home/esdata/snapshot /home/esdata/snapshot nfs4 ro,relatime,vers=4.2,rsize=1048576,wsize=1048576,namlen=255,hard,proto=tcp,timeo=600,retrans=2,sec=sys,clientaddr=192.168.10.26,local_lock=none,addr=192.168.10.141 0 0

PUT _snapshot/local
{
  "type": "fs",
  "settings": {
    "location": "snapshot",
    "readonly": true
  }
}

GET /_snapshot/local/_current
{
  "snapshots" : [ ],
  "total" : 0,
  "remaining" : 0
}

The above is operating on the New Node.

[wajika]

On the new cluster, I deleted the previously registered repo and created a new type:url repo. The same operation, the prompt I got is still [local:snapshot_2] snapshot does not exist

PUT _snapshot/local
{
  "type": "url",
  "settings": {
    "url": "file:/home/esdata/snapshot"
  }
}


POST /_snapshot/local/snapshot_2/_restore
{
  "error" : {
    "root_cause" : [
      {
        "type" : "snapshot_restore_exception",
        "reason" : "[local:snapshot_2] snapshot does not exist"
      }
    ],
    "type" : "snapshot_restore_exception",
    "reason" : "[local:snapshot_2] snapshot does not exist"
  },
  "status" : 500
}

[wajika]

I noticed that other people's solutions are to directly overwrite the data of nodeA to the data of nodeB, but I saw in the elastic documentation that nodeB can only mount the NFS directory in read-only mode. Is my understanding wrong?

[wajika]

I installed the cluster of version 7.16.3. I replaced the snapshot directory generated by version 7.12.0 with the directory created in advance by version 7.16.3, and then prompted that "[local:beifen/WMgGeuykR-KMhD96mwcnRw]" was missing. In this case, it should be It is not possible to directly use the method of directory replacement.

I want a more professional answer.