Migrate internal indices | Elasticsearch Add-On for Heroku Documentation

› › ›

« Migrate your Elasticsearch data Configuring your deployment »

Migrate internal indicesedit

When you migrate your Elasticsearch data into a new infrastructure you may also want to migrate your Elasticsearch internal indices, specifically the .kibana index and the .security index.

There are two ways to migrate the internal Elasticsearch indices:

Reindex the indices from a remote cluster.
Restore the indices from a snapshot.

To reindex internal indices from a remote cluster, you can follow the same steps that you use to reindex regular indices when you migrate your Elasticsearch data indices.

To restore internal indices from a snapshot, the procedure is a bit different from migrating Elasticsearch data indices. Use these steps to restore internal indices from a snapshot:

On your old Elasticsearch cluster, choose an option to get the name of your snapshot repository bucket:
```
GET /_snapshot
GET /_snapshot/_all
```

Get the snapshot name:

GET /_snapshot/NEW-REPOSITORY-NAME/_all

The output for each entry provides a "snapshot": value which is the snapshot name.

  {
  "snapshots": [
    {
      "snapshot": "scheduled-1527616008-instance-0000000004",

To restore internal Elasticsearch indices, you need to register the snapshot repository in read-only mode. To do so, first add the authentication information for the repository to the Elasticsearch Add-On for Heroku keystore, following the steps for AWS S3, Google Cloud Storage, or Azure Blog storage.
To register a read-only repository, open the Elasticsearch API console or the Kibana Dev Tools page and run the Read-only URL repository API call.
Once the repository has been registered and verified, you are ready to restore the internal indices to your new cluster, either all at once or individually.
- Restore all internal indices
  
  Run the following API call to restore all internal indices from a snapshot to the cluster:
```
POST /_snapshot/repo/snapshot/_restore
{
  "indices": ".*",
  "ignore_unavailable": true,
  "include_global_state": false,
  "include_aliases": false,
  "rename_pattern": ".(.+)",
  "rename_replacement": "restored_security_$1"
}
```
- Restore an individual internal index
  
  When restoring internal indices, ensure that the include_aliases parameter is set to false. Not doing so will make Kibana inaccessible. If you do run the restore without include_aliases, the restored index can be deleted or the alias reference to it can be removed. This will have to be done from either the API console or a curl command as Kibana will not be accessible.
  
  Run the following API call to restore one internal index from a snapshot to the cluster:
```
POST /_snapshot/repo/snapshot/_restore
{
  "indices": ".kibana",
  "ignore_unavailable": true,
  "include_global_state": false,
  "include_aliases": false,
  "rename_pattern": ".(.+)",
  "rename_replacement": "restored_security_$1"
}
```
  Next, the restored index needs to be reindexed into the internal index, as shown:
```
POST _reindex
{
  "source": {
    "index": "restored_kibana"
  },
  "dest": {
    "index": ".kibana"
  }
}
```

Your internal Elasticsearch index or indices should now be available in your new Elasticsearch cluster. Once verified, the restored_* indices are safe to delete.

« Migrate your Elasticsearch data Configuring your deployment »