DMaaS

Data Migration as a Service(DMaaS) is an offering from MayaData, available on both platforms - DirectorOnline and Director OnPrem. DMaaS is a solution for data agility and data motion in the Kubernetes platform, providing utilities and workflows to migrate Kubernetes stateful workloads along with their persistent data from anywhere to anywhere, be it on-premise or across clouds. It is application-aware, so the application state remains consistent before and after migration. Thus, it helps the user in backup and disaster recovery of the application along with the data, with the minimized time taken for restoring back to the normal operation mode.

Backup and Restore is supported for all volumes except LocalPV hostPath.

Scenarios where user can use DMaaS:

1. Migration of Kubernetes stateful workloads across clouds or between on-premise and cloud.
2. Disaster recovery
3. Application upgrades
4. Kubernetes upgrades
5. Data Backup and Recovery
6. Data Protection

DMaaS Operational Mode:

DMaaS backup agent will be deployed on both the source and destination clusters once these clusters are in Premium or Evaluation plan. The backup agent on the source cluster uses the provider APIs to send the backup to the object storage. The backup agent on the destination cluster retrieves the backup from the object storage and deploys the application from the backed-up configuration whenever it demands the backup data. The data motion can be scheduled for both the upload and download so that periodic incremental backups can be uploaded from the source and the destination can be kept in sync with the source.

Prerequisites

1. DirectorOnline Account with a subscription of Premium plan or Evaluation plan.
2. AWS S3, GCP or MinIO credentials.
3. Same StorageClass configuration and Service Account should exist on the destination cluster.
4. All app-specific resources like service, PVC, shared resources like Service Account, ClusterRoles, etc. should have a common label.
5. Kubernetes Version 1.12 or above is recommended.

Setting up the Environment:

1. Connect clusters to DirectorOnline. To know the steps for connection Click Here.
2. Add AWS or GCP or MinIO credentials for the first time with correct details of your account with a title name. Credentials with title name can be selected for subsequent backup schedules and restores.
3. Schedule the backup period using the AWS S3, GCP or MinIO credentials, select proper time interval and region (in case of AWS S3) while creating Data Motion schedule. Provide backup schedule by selecting the time slot.
4. Verify the list of Backup from the DMaaS page or from corresponding application page.
5. Restore the application to same cluster or different cluster from the selected backup.

Setting up Data Motion Schedule

1. DMaaS helps the user to do the Data Motion of the application data to a third party storage location like AWS or GCP or MinIO in a scheduled manner. User has to schedule the period from the DMaaS section present in the Application page.
2. With the current implementation, AWS S3,  MinIO and GCP are the third party storage location where the backup of the application is getting stored.

Steps to schedule the backup transfer of the corresponding application.

NOTE: Schedule or Backups older than 1 year will not be restored. This happens mainly because Velero deletes the base backup from Object Store (S3/GCP/MinIO) after the set ttl period, which in this case is 1 year.

STEP 1:
Clicking on the application on the left sidebar redirects to a page where you can select the application. By clicking on the application name you will be redirected to the corresponding application page.

STEP 2:
Select DMaaS on the selection box. Click on New schedule button.



STEP 3:
Next, a backup schedule needs to be created.
Follow the below mentioned steps to add credentials and schedule a backup.

STEP 4:
Next, select the time interval as desired(This is the interval after which backup will occur).

STEP 5:

NOTE: If the schedule is being created for cStor volume select the option as shown in the image below. 



If the schedule is being created for volumes other than cStor, the application pod will get restarted because restic annotations are added to the deployments.

Then, schedule the backup using Schedule now button.



STEP 6:
As soon as the DMaaS agents are in running state, backup starts automatically.
Once backup is completed you can see an auto generated schedule name on your screen.

Restoring the application Restoration of the application can be done in two ways: