Skip to main content

Création d’un réplica à haute disponibilité

Dans une configuration active/passive, l’appliance réplica est une copie redondante de l’appliance primaire. Si l’appliance primaire échoue, le mode haute disponibilité permet au réplica d’agir comme appliance principale, ce qui permet une interruption minimale du service.

Note

There is a maximum of 8 high availability replicas (both passive and active/geo replicas, as well as repository cache instances) allowed for GitHub Enterprise Server.

Creating a high availability replica

  1. Set up a new GitHub Enterprise Server appliance on your desired platform. The replica appliance should mirror the primary appliance's CPU, RAM, and storage settings. We recommend that you install the replica appliance in an independent environment. The underlying hardware, software, and network components should be isolated from those of the primary appliance. If you are a using a cloud provider, use a separate region or zone. For more information, see "Setting up a GitHub Enterprise Server instance."

  2. Ensure that the new appliance can communicate with all other appliances in this high availability environment over ports 122/TCP and 1194/UDP. For more information, see "Network ports."

  3. In a browser, navigate to the new replica appliance's IP address and upload your GitHub Enterprise license.

  4. Set an admin password that matches the password on the primary appliance and continue.

  5. Click Configure as Replica.

  6. Under "Add new SSH key", type your SSH key.

  7. Click Add key.

  8. Connect to the replica appliance's IP address using SSH.

    ssh -p 122 admin@REPLICA_IP
    
  9. To generate a key pair for replication, use the ghe-repl-setup command with the primary appliance's IP address and copy the public key that it returns.

    ghe-repl-setup PRIMARY_IP
    
  10. To add the public key to the list of authorized keys on the primary appliance, browse to https://PRIMARY-HOSTNAME/setup/settings and add the key you copied from the replica to the list.

  11. To verify the connection to the primary and enable replica mode for the new replica, run ghe-repl-setup again.

    ghe-repl-setup PRIMARY_IP
    
  12. To start replication of the datastores, use the ghe-repl-start command.

    ghe-repl-start
    

    Warning: ghe-repl-start causes a brief outage on the primary server, during which users may see internal server errors. To provide a friendlier message, run ghe-maintenance -s on the primary node before running ghe-repl-start on the replica node to put the appliance in maintenance mode. Once replication starts, disable maintenance mode with ghe-maintenance -u. Git replication will not progress while the primary node is in maintenance mode.

  13. To verify the status of each datastore's replication channel, use the ghe-repl-status command.

    ghe-repl-status
    

Creating geo-replication replicas

This example configuration uses a primary and two replicas, which are located in three different geographic regions. While the three nodes can be in different networks, all nodes are required to be reachable from all the other nodes. At the minimum, the required administrative ports should be open to all the other nodes. For more information about the port requirements, see "Network ports."

The latency between primary and replica nodes must be less than 70 milliseconds. We don't recommend configuring a firewall between the nodes' networks. If latency is more than 70 milliseconds, we recommend cache replica nodes instead. For more information, see "Configuring a repository cache."

  1. Create the first replica the same way you would for a standard two node configuration by running ghe-repl-setup on the first replica.

    (replica1)$ ghe-repl-setup PRIMARY_IP
    (replica1)$ ghe-repl-start
    
  2. Create a second replica and use the ghe-repl-setup --add command. The --add flag prevents it from overwriting the existing replication configuration and adds the new replica to the configuration.

    (replica2)$ ghe-repl-setup --add PRIMARY_IP
    (replica2)$ ghe-repl-start
    
  3. By default, replicas are configured to the same datacenter, and will now attempt to seed from an existing node in the same datacenter. Configure the replicas for different datacenters by setting a different value for the datacenter option. The specific values can be anything you would like as long as they are different from each other. Run the ghe-repl-node command on each node and specify the datacenter.

    On the primary:

    (primary)$ ghe-repl-node --datacenter [PRIMARY DC NAME]
    

    On the first replica:

    (replica1)$ ghe-repl-node --datacenter [FIRST REPLICA DC NAME]
    

    On the second replica:

    (replica2)$ ghe-repl-node --datacenter [SECOND REPLICA DC NAME]
    

    Tip: You can set the --datacenter and --active options at the same time.

  4. An active replica node will store copies of the appliance data and service end user requests. An inactive node will store copies of the appliance data but will be unable to service end user requests. Enable active mode using the --active flag or inactive mode using the --inactive flag.

    On the first replica:

    (replica1)$ ghe-repl-node --active
    

    On the second replica:

    (replica2)$ ghe-repl-node --active
    
  5. To apply the configuration, use the ghe-config-apply command on the primary.

    (primary)$ ghe-config-apply
    

Configuring DNS for geo-replication

Configure Geo DNS using the IP addresses of the primary and replica nodes. You can also create a DNS CNAME for the primary node (e.g. primary.github.example.com) to access the primary node via SSH or to back it up via backup-utils.

For testing, you can add entries to the local workstation's hosts file (for example, /etc/hosts). These example entries will resolve requests for HOSTNAME to replica2. You can target specific hosts by commenting out different lines.

# <primary IP>      HOSTNAME
# <replica1 IP>     HOSTNAME
<replica2 IP>     HOSTNAME

Further reading