This procedure explains how to migrate a Flow Production Tracking hosted site to a local flow production tracking instance.

Requirements

This procedure assumes that:

• For the media, the host running the flow production tracking-app containers can access AWS S3 end-points.
• Your media storage is properly configured and has enough storage capacity to store all your media.
• Flow Production Tracking Enterprise (SE) Docker is already running and properly configured.

Step 0. Create a ticket with flow production tracking Support

Our support team is there to help you with the migration process. If you have not already done so, please open a ticket with Flow Production Tracking Support and they will guide you during that process.

Step 1. Restore the database

Download the DB snapshot from S3 using the credentials and path we supply. You will need to install aws-cli and configure it with the credentials we supply. We will provide the necessary <bucket> and <prefix> values alongside the credentials. 

 aws s3 cp s3://<bucket>/<prefix>/db_backup /opt/flow production tracking/se/production/db_backup --recursive 

Restore the hosted database with on your flow production tracking instance.

docker-compose stop app emailnotifier  
docker-compose exec dbops dropdb flow production tracking  
docker-compose exec dbops createdb flow production tracking  
docker cp <BACKUP_NAME> $(docker-compose ps -q dbops):/db_backup/  
docker-compose exec dbops pg_restore --no-owner --verbose -Fc --no-acl --jobs 4 --dbname flow production tracking /db_backup/<backup_file>.dump  
docker-compose run --rm app rake admin:init_database  
docker-compose run --rm app rake admin:reset_flow production tracking_admin_password  
docker-compose start app emailnotifier 

Step 2. Configure access to AWS S3

This step is optional. Configuring S3 will allow you to access the media you previously injected in your hosted instance. Your local Flow Production Tracking Cluster will be functional without the media, but your Flow Production Tracking instance won’t be able to display thumbnails, filmstrip, and movies that were added while your hosted instance was used in production.

Configure flow production tracking to use AWS S3

 vi /opt/flow production tracking/se/production/services.yml 

Use this template and replace ACCESS_KEY_ID and SECRET_KEY_ID with the values provided by Flow Production Tracking Support. If you don’t have these keys, please ask for them in your Support ticket.

aws:  
  auth:  
    app:  
      access_key: ACCESS_KEY_ID  
      secret_key: SECRET_KEY_ID  
  s3:  
    sg-media-usor-01:  
      region: us-west-2  
    sg-media-tokyo:  
      region: ap-northeast-1  
    sg-media-ireland:  
      region: eu-west-1  

For all your flow production tracking-app container(s), map the services.yml by adding it as a volume in the docker-compose.yml file.

app:  
  volumes:  
    - ./media:/media  
    - /opt/flow production tracking/se/production/services.yml:/var/rails/flow production tracking/current/config/services.yml 

Restart app container

docker-compose up -d app 

You should now have be able to see the media from S3 in your local Flow Production Tracking instance, typically at http://<your_site>.com.

Download the media

Only for Flow Production Tracking SE Docker v7.4.x and later.

docker-compose run --rm app script/sync_s3_media.rb 

This will run for hours if the hosted site has a lot of media.

Remove S3 access

When all the media has been downloaded, you can remove the service.yml file and remove the host access to S3.

Troubleshooting

There is a 403 error in Flow Production Tracking while browsing media after configuring AWS S3.

If you get 403 errors on the web page, double check the server time. It must be up-to-date to the minute; the AWS API is really picky about this. A classic option is to use ntpdate.

sudo yum install ntpdate -y && sudo ntpdate pool.ntp.org 

The media downloading script failed. Can it be run again?

Yes, this script is idempotent. If it is rerun after failure, media that has already been downloaded will be skipped.

What is happening during the download process?

Files are downloaded locally on your media server, and the reference to the file is updated in the flow production tracking database. Files are not removed from S3 during that process.

Database Error

When you see this error

dropdb: database removal failed: ERROR:  database "flow production tracking" is being accessed by other users  
DETAIL:  There is 1 other session using the database.

Please follow this:

docker-compose run dbops sh  
# psql  
psql (9.6.7, server 9.6.6)  
Type "help" for help.   
flow production tracking=# SELECT pg_terminate_backend(pg_stat_activity.pid)  
    FROM pg_stat_activity  
    WHERE pg_stat_activity.datname = 'flow production tracking'  
      AND pid <> pg_backend_pid();