Backup and Restore (Docker Swarm)

For Docker Swarm based deployments, Cisco Provider Connectivity Assurance provide solutions for two classes of issues:

1. Targeted application issues

2. VM loss due to infrastructure issues

Targeted Backup and Restoration

Configuration data is automatically backed up daily to an S3 provider deployed as part of the solution. This S3 provider provides redundancy of storage on the allocated disk partitions.

The backups are used in case of specific database corruption or accidental data deletions scenarios. MOPs for restoration would be provided in response to specific TAC cases.

Generically the MOPs would involve:

1. Exporting backup files from minIO using “mc” CLI

2. Moving files to the VMs running service to be restored

3. Stopping/removing service and replacing files OR leveraging database specific restoration tools

4. Running a redeployment script to replace stopped/removed services and restart dependent services

Infrastructure Issue Requiring Rebuild

For issues involving the loss of entire VMs and their data partitions, the application configuration backups are not the most expedient method to return to service. For these catastrophic loss scenarios, we recommend taking snapshots of the following directories:

All VM’s in the swarm cluster:

1. /var/lib/docker

Leader VM only:

1. /var/lib/docker

2. /etc/aod

3. /home/admin

Restoration from VM loss would leverage these filesystem backups to get the services running again based on MOPs provided by Cisco TAC.