Host external ID and deployment script

When configuring a new host and registering it to PrivX using deploy script, the host is assigned a unique external ID. For cloud hosts, this external ID is based on instance ID on cloud instance metadata. Instance IDs are automatically used, if you use --aws, --google-cloud, --azure or --openstack deployment flags.

For on-prem hosts and hosts configured with deploy script but without cloud provider flag, the host's machine-id is used instead.

If you're deploying cloned virtual machines or on-prem hosts using machine-id, make sure your machine id's are unique before running the deploy script. The same thing applies, if running PrivX instances on on-prem hosts. Running HA instances on hosts with duplicate IDs might cause issues with license activations.

📘

Note

You can check your host's machine ID from /etc/machine-id, or by running:

# dmidecode --string system-uuid

For more information about machine IDs, see https://www.man7.org/linux/man-pages/man5/machine-id.5.html

Re-running a deploy script on a host will update the existing host in PrivX host store. Re-running the deployment script requires host's 'deployable' flag to be enabled first for security reasons. If the host does not exist, it will be created. If the host was added using PrivX host directory host scanning feature, the host will have the same external_id as when deploying it with a deployment script and will be updated accordingly.

You should not configure multiple identical host directories which see the same instances. If this is necessary for some reason, use host tagging and "FETCH HOSTS WITH TAG" -feature in PrivX host directory to filter out unwanted hosts.

Hosts added manually via PrivX UI are considered 'local hosts' and have no external ID. Those instances cannot be redeployed via deployment script and need to be configured manually or via automation tools instead.


Did this page help you?