Skip to main content
Version: 2.2

Install data transfer agents

Data transfer agents let you scale beyond the limitations of a single host by sharing the workload of transferring data across additional hosts with access to your source. Data transfer agents accelerate data transfer by removing network, memory, and CPU bottlenecks.

This means Data Migrator can be scaled to the capacity of your wide area network or another limitation, such as the data transfer capability of your storage environment.

You can add an unlimited number of data transfer agents to Data Migrator.

caution

Data transfer agents are not supported with local fs sources.

note

Data transfer agents assist with moving data but aren't involved in scaling metadata migrations, which are typically under a much lower load and don't experience bottlenecks from being deployed on a single host.

info

If you decide to use data transfer agents, agents become responsible for transferring the data to your target filesystem(s). Data Migrator doesn't transfer the data. If you stop the agents, your data isn't migrated. If you want to stop using agents, and use Data Migrator to move the data, you must deregister your agents.
See the section Remove an agent.

If you want to manage traffic, do so on standalone instances of Data Migrator using bandwidth limits.

info

Older operating systems (CentOS 6) may not have the required libraries to provide secure channel communication. See the following Knowledge base article to disable secure channel communication.

Prerequisites​

  • You are the system administrator.

  • Additional hosts are deployed on your network that can access the source storage environment.

  • If you're transferring data from a Hadoop environment, you have the default name hdfs for your system user and system group.
    See Configure system users.

  • Port 1433 is open between the host running Data Migrator and all hosts running data transfer agents.
    See Network requirements.

note

The agent must have the same version as Data Migrator. If the agent version isn't the same, the agent status isn't recognized as valid.

For recommended machine specifications, see the Installation Prerequisites.

Install an agent​

Install each agent on a separate host with client access to the source filesystem.

  1. Download either the RPM or the Debian installer for your selected host. Links to the applicable version of the data transfer agent are available if you select Add a data transfer agent on the data transfer agent page in the WANdisco UI.

    • RPM
      curl -L -o livedata-migrator-data-agent-installer.sh https://docs-router.wandisco.com/router?url=/data-transfer-agents/download?version=1.25.3-rpm
    • Debian
      curl -L -o livedata-migrator-data-agent-installer.sh https://docs-router.wandisco.com/router?url=/data-transfer-agents/download?version=1.25.3-deb

  2. Make the installer executable: 

    chmod +x livedata-migrator-data-agent-installer.sh

  3. Run the installation command as the root (or sudo) user: 

    ./livedata-migrator-data-agent-installer.sh
info

By default, the agent will be installed using the hdfs user and hdfs group. To install the agent with a non-default system user see Custom user installation below.

Custom user installation​

To install the agent using a custom user or a custom user group, use the following command:

Agent install with custom user
./livedata-migrator-data-agent-<version>_installer.sh -- --user <custom user> --group <custom group>

Authentication​

A successful installation prints the authentication token in your terminal for Data Migrator to authenticate to the agent.

The format of the generated token is: 

----- BEGIN AGENT TOKEN

----<example_token_text>----

END AGENT TOKEN -----

The authentication token is generated with the hostname of the system in the /etc/hostname file. The hostname may be different to the hostname of the node on the network. To ensure the node is valid when added to the UI, the generator includes a script - gen-certs-and-token.sh - that allows you to create custom hostnames and ports manually.

Run the script as described in steps 1 to 3 above.

Custom user authentication​

If you are a custom user, run the following commands: 

  1. chown <custom_user>:<custom_group> gen-certs-and-token.sh
  2. sudo -u <custom_user> ./gen-certs-and-token.sh --force

The output details the directories in which agent files are stored. See the output example below.

Example output from a successful installation
   
Token Generator

Certificates and keys will be stored in /opt/wandisco/livedata-migrator-data-agent/certs,
   Keystore = keystore.p12,
   Data Agent Host = example01-vm3.bdauto.wandisco.com,
   Data Agent Server Port = 1433
   
Connection Token will be stored in /opt/wandisco/livedata-migrator-data-agent/connection_token,
Security config file is /etc/wandisco/livedata-migrator-data-agent/security.properties,
Registration request file is /opt/wandisco/livedata-migrator-data-agent/reg_data_agent.json

(Re)start the livedata-migrator-data-agent service and call the livedata-migrator API to (re)register an agent:
curl -XPOST -H "Content-Type: application/json" -d @/opt/wandisco/livedata-migrator-data-agent/reg_data_agent.json http://host:18080/scaling/dataagents/
Agent connection token:
----- BEGIN AGENT TOKEN -----
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
----- END AGENT TOKEN -----

WANdisco LiveData Migrator Data Agent installed successfully.
note

You don't need to start the livedata-migrator-data-agent service.

Copy the authentication token from the output in your terminal.

Alternatively, you can find the generated token in /opt/wandisco/livedata-migrator-data-agent/reg_data_agent.json.

Configure the agent with the UI​

Add an agent with the UI​

To add an agent with the UI, follow these steps:

  1. Select your Data Migrator product from the Products list in the dashboard.
  2. Under Filesystems and Agents, select Data Transfer Agents.
  3. Select Add a data transfer agent.
  4. Follow the steps in the UI to install the agent in your terminal. See Install an agent.
  5. Paste the authentication token from the output in your terminal into the Authentication Token field.
    Data Migrator derives the hostname and the port number from this token and populates the corresponding fields in the UI automatically.
  6. Select Save.

Check the status of an agent with the UI​

To check if the status of the connection to an agent is healthy:

  1. From the Dashboard, select a product under Products.
  2. Under Filesystems and Agents, select Data Transfer Agents. The Data Transfer Agents page lists whether the agents are healthy or unhealthy under the Status column.

Check the progress of an agent migrating data with the UI​

To check the progress of an agent migrating data:

  1. From the Dashboard, select a product under Products.
  2. Under Filesystems and Agents, select Data Transfer Agents.
  3. Select an agent from the Data Transfer Agents page.
  4. On the Data Transfer Agent Progress panel, view the following:
    • Data transferred in the past 24 hours (up to the last hour)
    • Data transferred in the past week

Metrics appear only after the agent begins to transfer data. You can also view bandwidth usage for your configured agents. An overview of migrated data is available on the Metrics tab for the product.

Set up email notifications​

To get notified by email about the status of agents, go to the Email Notifications page to set up these alerts.
For more information, see Configure email notifications.

Remove an agent with the UI​

To remove an agent, follow these steps:

  1. From the Dashboard, select a product under Products.
  2. Under Filesystems and Agents, select Data Transfer Agents.
  3. Select the agent from the list that you want to remove.
  4. In the panel Data Transfer Agent Connection, select Remove.

The agent no longer appears in the list.

Uninstall an agent​

To uninstall an agent, use one of the following commands:

Uninstall the RPM package (.rpm)
yum remove -y livedata-migrator-data-agent

Or

Uninstall the Debian package (.deb)
apt-get purge -y livedata-migrator-data-agent

(Optional) Delete all related agent directories:

Delete all related agent directories
rm -rf /etc/wandisco/livedata-migrator-data-agent /var/log/wandisco/livedata-migrator-data-agent /var/run/livedata-migrator-data-agent /opt/wandisco/livedata-migrator-data-agent

Upgrade an agent​

To upgrade to the latest version of the data transfer agent, download and run a new installer (see install an agent). After upgrading, the agent version is updated automatically in the UI.

If you've got only one agent, there's a short period of time when the agent version and the Data Migrator version don't match. This may temporarily affect your filesystems or migrations until the versions are matched. To prevent any disruption to migrations, stop all data migrations before you upgrade. Resume stopped data migrations manually after the versions match.
For more information on stopping and resuming data migrations manually, see Bulk actions.

If you've got multiple agents, don't upgrade them all at once. When upgraded agents match the Data Migrator version and resume transferring data, you can continue to upgrade the other agents. This way, upgrading to a newer version of your agents won't affect your filesystems or migrations. Any migrations that are in progress continue transferring data as normal.

After upgrading, old agent values in the vars.env file are replaced by new ones, and new properties are added to the end of the file.

The minimum support version of Data Migrator with data transfer agents is version 1.22.0.

Learn more​

Configure data transfer agent properties
Secure communication
Troubleshooting