snapshot_restore_exception migrating ElasticSearch 7 indices between servers - elasticsearch

Question

We want to migrate an ElasticSearch v7 index from server 1 to server 2.
How can we copy and restore a snapshot from one server to another?
The offical documentation seems to assume both servers share the same physical storage. These servers do not "know" each other and have completely seperate disks. In our case we:
Create the same repository foo on both servers
Run a snapshot foo1 on server 1
Create the /snapshots/foo directory on server 2
Verify the repo on server 2
Run a snapshot footest on server 2 to verify files are written to /snapshots/foo
Copy across the /snapshots/foo directory from server 1 to server 2
Run the restore of foo1:
POST /_snapshot/foo/foo1/_restore
{
"indices": "foo"
}
The error is:
{
"error": {
"reason": "[foo:foo1] snapshot does not exist",
"root_cause": [
{
"reason": "[foo:foo1] snapshot does not exist",
"type": "snapshot_restore_exception"
}
],
"type": "snapshot_restore_exception"
},
"status": 500
}
When we check the repo:
GET /_snapshot/foo/*
We can see only one snapshot (which we created on server 2). We cannot see the snapshot we copied across from server 1.
So how does one transfer a snapshot to server 2?

Answer 1

#Marc there are few options for you to copy data from one server to another in terms of indices:
Configure CCR (cross cluster replication) for your index and let it sync between the two servers (which can be on separate machines. network/ etc... they should just be able to communicate)
You can have a single repository created which can be accessed by both servers (using shared IAM user or role assuming it's on AWS). You can then take snapshot of your index from server1 and restore the snapshot in server2.
Not recommended but you can copy ${path.data} files from server1 to ${path.data} in server2.
However, the main question here would be: are your server1 and server2 running on same ES version ?