Storage
Can existing storage be used in the containerized environment?
Yes. Because of the hybrid solution chosen, storage from existing environments can be reused in the container environment.
There are two configuration options:
Shared storage: Set up entirely on the host machines — supported via bind mounts set up in the Docker Swarm config.
Hybrid approach: NFS (or other supported protocol) server set up outside of the Docker Swarm configuration. Each container will be configured to connect to the NFS server using Docker Swarm volumes.
Note
Shared storage set up entirely inside of Docker alongside the application container is possible but is not recommended and not as performant as the recommended configuration options.
What option is there for environments without network storage?
If a customer does not use network attached storage for /data and /localdata, that means the NFS shares are likely hosted on the Application and/or Analytics hosts. As is the case whenever provisioning new hardware/VMs for a Brainspace instance, the data that is in /data and /localdata on the old machines needs to be copied to the new machines. This can be done offline ahead of time, in preparation for the migration.
Once the new application host is ready, rsync can be used to start copying data from the old machines to the new application host. When the Brainspace 7.0 installation is performed, BRS dataset data will already be prepared in the new location.