This repostory contains:
- Sample Cloud foundry Java Spring Webapp
- BBR restore script for restoring HA deployment's online S3 backup artifacts to any HA Postgres deployment
- Java 8
- Maven 3.0 or newer
- Cloud Foundry CLI
To build the application, run the following command from the postgres-consumer-app directory:
mvn clean install && mv target/postgres-demo-0.0.1-SNAPSHOT.jar ./
To deploy the application, first edit the APP_NAME,SVC_INSTANCE_NAME and SSL_MODE in the postgres-consumer-app/manifest.yml file.
The SSL_MODE values are listed here.
Then run the following command from the postgres-consumer-app directory:
cf push
This will create, push and bind the application to the service instance you mentioned in manifest.yaml file.
This script is useful in only this case. Although running it in any other case will also work, it will be simply a regular BBR restore.
The backup artifacts from HA service instance need a modification to be used with bbr restore. As you can see in the directory structure for HA backup artifacts, there are 3 metadata files generated and saved, 1 for each postgres instance. To start your restore, first download the whole folder corresponding to your HA service deployment, as described in the structure section. BBR needs only 1 metadata file to be used during restore, which is essentially a union of all 3 files. The following script does all of this.
Place the file ha-bbr-restore.sh inside the top level service-instance_ folder downloaded from your configured S3 bucket for backups. After this, only keep the timetamp folders that you want. You can delete all other timestamp folders from all VM instance folders. Then:
cp ./ha-bbr-restore.sh downloaded-s3-backup-folder/
cd downloaded-s3-backup-folder/
./ha-bbr-restore.sh
After creating necessary files, this will run regular BBR restore. No additional steps need to be taken.