Table of Contents

Migrating from Elasticsearch to OpenSearch

From DataMiner 10.4.0 [CU2]/10.4.4 onwards, a tool is available that allows you to migrate from Elasticsearch 6.8.22 to OpenSearch 2.11.1.

To use this tool, follow the instructions below:

  1. Stop all Agents in the DMS.

  2. Take a snapshot of the Elasticsearch 6.8.22 cluster.

  3. Copy the snapshot to an Elasticsearch 7.10.0 cluster and restore it.

  4. Run the re-indexing tool and take a snapshot.

  5. Copy the snapshot with the re-indexed data to an OpenSearch 2.11.1 cluster and restore it.

  6. Reconfigure DB.xml on every Agent.

  7. Restart all DataMiner Agents in the DMS.

Caution

To prevent data loss, all Agents of the DMS must be stopped during this procedure. They must not be started up again until the migration is completed.

Tip

See also: Taking a snapshot of one Elasticsearch cluster and restoring it to another

Take a snapshot of the Elasticsearch 6.8.22 cluster

Using Kibana, you can take a snapshot in the following way:

  1. Check the path.repo configuration in elasticsearch.yml.

    This configuration should point to a shared file system location to which each node has access.

  2. Check the existing repositories by sending the following request:

    GET /_snapshot/_all

    • This will return a response containing information about all the repositories configured in Elasticsearch. In case the desired repository already exists, you can skip the next step.

    • Example response when a repository already exists:

      {
  "test_tepo_1" : {
    "type" : "fs",
    "settings" : {
      "location" : "/usr/share/elasticsearch/repo",
      "maxRestoreBytesPerSec" : "40mb",
      "readonly" : "false",
      "compress" : "true",
      "maxSnapshotBytesPerSec" : "40mb"
    }
  }
}

  3. Create the repository by sending the following request:

    PUT /_snapshot/<repo_name>
 {
     "type": "fs",
     "settings": {
     "location" : "<shared_repo_path>",
         "maxRestoreBytesPerSec" : "40mb",
         "readonly" : "false",
         "compress" : "true",
         "maxSnapshotBytesPerSec" : "40mb"
     }
 }

    Variables:

    • repo_name: A repository name of your choice.
    • shared_repo_path: The path to the shared folder where the snapshots will be stored.

  4. Take the snapshot by sending the following request:

    PUT /_snapshot/<repo_name>/<snapshot_name>
 {
   "indices": "dms*"
 }

    Variable:

    • snapshot_name: A snapshot name of your choice.

  5. Check the snapshot by sending the following request:

    GET /_snapshot/<repo_name>/<snapshot_name>/_status

    This request returns status information for a specific snapshot within a given repository. The state should be SUCCESS and all shards should be successful.

Restore the snapshot on an Elasticsearch 7.10.0 cluster

Using Kibana, you can restore the snapshot in the following way:

  1. Check the path.repo configuration in elasticsearch.yml.

    This configuration should point to a shared file system location to which each node has access.

  2. Check the existing repositories by sending the following request:

    GET /_snapshot/_all

    • This will return a response containing information about all the repositories configured in Elasticsearch. In case the desired repository already exists, you can skip the next step.

    • Example response when a repository already exists:

      {
  "test_tepo_1" : {
    "type" : "fs",
    "settings" : {
      "location" : "/usr/share/elasticsearch/repo",
      "maxRestoreBytesPerSec" : "40mb",
      "readonly" : "false",
      "compress" : "true",
      "maxSnapshotBytesPerSec" : "40mb"
    }
  }
}

  3. Create the repository by sending the following request:

    PUT /_snapshot/<repo_name>
 {
     "type": "fs",
     "settings": {
     "location" : "<shared_repo_path>",
         "maxRestoreBytesPerSec" : "40mb",
         "readonly" : "false",
         "compress" : "true",
         "maxSnapshotBytesPerSec" : "40mb"
     }
 }

  4. Restore the snapshot by sending the following request:

    POST /_snapshot/<repo_name>/<snapshot_name>/_restore

  5. Check the snapshot by sending the following request:

    GET /_snapshot/<repo_name>/<snapshot_name>/_status

    This request will return information about the status of the specified snapshot. The status should be "SUCCESS".

  6. Check the cluster health by sending the following request:

    GET /_cluster/health

    The status of the cluster should turn green after the restore.

Run the re-indexing tool and take a snapshot

  1. Open a terminal, and go to the folder containing the tool. By default, this is the folder C:\Skyline DataMiner\Tools\ReIndexElasticSearchIndexes:

    cd C:\Skyline DataMiner\Tools\ReIndexElasticSearchIndexes

  2. Run the tool with the following arguments:

    Argument Description
    -Node or -N The name of the node to be used for re-indexing (mandatory).
    Format: http(s)://127.0.0.1:9200 or http(s)://fqdn:9200
    -User or -U The user name, to be provided in case Elasticsearch was hardened.
    See Securing the Elasticsearch database
    -Password or -P The user password
    -DBPrefix or -D The database prefix, to be provided in case a custom database prefix is used instead of the default dms- prefix.
    If you do not provide a prefix, the default dms- will be used.
    -TLSEnabled or -T Whether or not TLS is enabled for this ElasticSearch database.
    Values: true or false. Default: false

  3. Take a snapshot of the re-indexed data by sending the following request:

    PUT /_snapshot/<repo_name>/<snapshot_name_reindexed>
 {
   "indices": "dms*"
 }

  4. Check the snapshot by sending the following request:

    GET /_snapshot/<repo_name>/<snapshot_name>/_status

    This request will return information about the status of the specified snapshot. The status should be "SUCCESS".

Restore the snapshot with the re-indexed data to an OpenSearch 2.11.1 cluster

  1. Check the path.repo configuration in opensearch.yml.

    This configuration should point to a shared file system location to which each node has access.

  2. Check the existing repositories by sending the following request:

    GET /_snapshot/_all

    • This will return a response containing information about all the repositories configured in Elasticsearch. In case the desired repository already exists, you can skip the next step.

    • Example response when a repository already exists:

      {
  "test_tepo_1" : {
    "type" : "fs",
    "settings" : {
      "location" : "/usr/share/elasticsearch/repo",
      "maxRestoreBytesPerSec" : "40mb",
      "readonly" : "false",
      "compress" : "true",
      "maxSnapshotBytesPerSec" : "40mb"
    }
  }
}

  3. Create the repository by sending the following request:

    PUT /_snapshot/<repo_name>
 {
     "type": "fs",
     "settings": {
     "location" : "<shared_repo_path>",
         "maxRestoreBytesPerSec" : "40mb",
         "readonly" : "false",
         "compress" : "true",
         "maxSnapshotBytesPerSec" : "40mb"
     }
 }

  4. Restore the snapshot by sending the following request:

    POST /_snapshot/<repo_name>/<snapshot_name_reindexed>/_restore

  5. Check the snapshot by sending the following request:

    GET /_snapshot/<repo_name>/<snapshot_name>/_status

    This request will return information about the status of the specified snapshot. The status should be "SUCCESS".

  6. Check the cluster health by sending the following request:

    GET /_cluster/health

    The status of the cluster should turn green after the restore.

Reconfigure DB.xml on every Agent

In case the OpenSearch nodes have been installed on a different machine than the Elasticsearch nodes you are migrating from, do the following on every Agent in your DMS:

  1. Navigate to the folder C:\Skyline DataMiner and open DB.xml.

  2. Locate the IP addresses or hostnames previously used by Elasticsearch.

    Tip

    For information on where these are located in DB.xml, see Indexing database settings.

  3. Replace these with the IP addresses of the OpenSearch nodes.

  4. Save the DB.xml file and exit.

When all DB.xml files have been reconfigured, you can restart all DataMiner Agents to finish the procedure.

Propose changes