OSS to Cluster Migration

The following guide has step-by-step instructions for migrating an InfluxDB OSS instance into and InfluxEnterprise cluster.

The process assumes that you already have a fully configured InfluxEnterprise cluster of three or more meta nodes and zero or more data nodes.

Please note that this migration process:

• Deletes all data from any data nodes that are already part of the InfluxEnterprise Cluster
• Will transfer all users from the OSS instance to the InfluxEnterprise Cluster*
• Will not transfer any users from the OSS instance to the InfluxEnterprise Web Console
• Requires downtime for writes and reads for the OSS instance

In addition, please refrain from creating a Global Admin user in the InfluxEnterprise Web Console before implementing these steps. If you’ve already created a Global Admin user, contact support.

Modify the /etc/hosts file

Add the IP and hostname of the OSS instance to the /etc/hosts file on all nodes in the InfluxEnterprise Cluster. Ensure that all cluster IPs and hostnames are also in the OSS instance’s /etc/hosts file.

Note: All node hostnames must be completely resolvable by all other nodes in the cluster. If you have DNS already setup in your environment, then this step can be skipped.

For all existing InfluxEnterprise data nodes:

1. Remove the node from the InfluxEnterprise Cluster

From a meta node in your InfluxEnterprise Cluster, enter:

influxd-ctl remove-data <data_node_hostname>:8088

2. Delete any existing data

On each data node that you dropped from the cluster, enter:

sudo rm -rf /var/lib/influxdb/{meta,data,hh}

3. Create new directories

On each data node that you dropped from the cluster, enter:

sudo mkdir /var/lib/influxdb/{data,hh,meta}

To ensure the file permissions are correct please run:

sudo chown -R influxdb:influxdb /var/lib/influxdb

For the OSS Instance:

1. Stop all writes to the OSS Instance

2. Stop the influxdb service on the OSS Instance

On sysvinit systems, use the service command:

sudo service influxdb stop

On systemd systems, use the systemctl command:

sudo systemctl stop influxdb

Double check that the service is stopped (the following should return nothing):

ps ax | grep influxd

3. Remove the OSS package

On Debian/Ubuntu systems:

sudo apt-get remove influxdb

On RHEL/CentOS systems:

sudo yum remove influxdb

4. Update the binary

Note: This step will overwrite your current configuration file. If you have settings that you’d like to keep, please make a copy of your config file before running the following command.

Ubuntu & Debian (64-bit)

wget https://dl.influxdata.com/enterprise/releases/influxdb-data_1.2.4-c1.2.5_amd64.deb
sudo dpkg -i influxdb-data_1.2.4-c1.2.5_amd64.deb

RedHat & CentOS (64-bit)

wget https://dl.influxdata.com/enterprise/releases/influxdb-data-1.2.4_c1.2.5.x86_64.rpm
sudo yum localinstall influxdb-data-1.2.4_c1.2.5.x86_64.rpm

5. Update the configuration file

In /etc/influxdb/influxdb-meta.conf, set:

• hostname to the full hostname of the meta node
• registration-enabled in the [enterprise] section to true
• registration-server-url in the [enterprise] section to the full URL of the server that will run the InfluxEnterprise web console. You must fully specify the protocol, IP or hostname, and port. Entering the IP or hostname alone will lead to errors.

• license-key in the [enterprise] section to the license key you received on InfluxPortal OR license-path in the [enterprise] section to the local path to the JSON license file you received from InfluxData. The license-key and license-path settings are mutually exclusive and one must remain set to the empty string.

  # Hostname advertised by this host for remote addresses.  This must be resolvable by all
  # other nodes in the cluster
  hostname="<enterprise-meta-0x>" 
  
  [enterprise]
  # Must be set to true to use the Enterprise Web UI
  registration-enabled = true 
  
  # Must include the protocol (http://)
  registration-server-url = "http://<web-console-server-IP>:3000" 
  
  # license-key and license-path are mutually exclusive, use only one and leave the other blank
  license-key = "<your_license_key>" # Mutually exclusive with license-path
  
  # license-key and license-path are mutually exclusive, use only one and leave the other blank
  license-path = "/path/to/readable/JSON.license.file" # Mutually exclusive with license-key
  

6. Start the data node

On sysvinit systems, use the service command:

sudo service influxdb start

On systemd systems, use the systemctl command:

sudo systemctl start influxdb

7. Add the node to the cluster

From a meta node in the cluster, run:

influxd-ctl add-data <data-node-hostname>:8088

You should see:

Added data node y at data-node-hostname:8088

Note: it may take a few minutes before the existing data become available in the cluster.

Final steps

1. Add any data nodes that you removed from cluster back into the cluster

From a meta node in the InfluxEnterprise Cluster, run:

influxd-ctl add-data <the-hostname>:8088

Output:

Added data node y at the-hostname:8088

Finally verify that all nodes are now members of the cluster as expected:

influxd-ctl show

2. Rebalance the cluster

Increase the replication factor on all existing retention polices to the number of data nodes in your cluster. You can do this with ALTER RETENTION POLICY.

Next, rebalance your cluster manually to meet the desired replication factor for existing shards.