2019-11-29 22:17:47 -07:00
# Introduction
2019-12-01 21:49:48 -07:00
This project provides Docker images to periodically back up a PostgreSQL database to AWS S3, and to restore from the backup as needed.
2019-11-27 20:57:21 -07:00
# Usage
## Backup
```yaml
postgres:
2021-06-20 14:22:25 -04:00
image: postgres:13
2019-11-27 20:57:21 -07:00
environment:
POSTGRES_USER: user
POSTGRES_PASSWORD: password
2019-12-01 18:47:18 -07:00
pg_backup_s3:
2021-06-20 14:22:25 -04:00
image: eeshugerman/postgres-backup-s3:13
2019-11-27 20:57:21 -07:00
environment:
2022-04-15 10:30:39 -07:00
SCHEDULE: '@weekly ' # optional
BACKUP_KEEP_DAYS: 7 # optional
PASSPHRASE: passphrase # optional
2019-11-27 20:57:21 -07:00
S3_REGION: region
S3_ACCESS_KEY_ID: key
S3_SECRET_ACCESS_KEY: secret
S3_BUCKET: my-bucket
S3_PREFIX: backup
2020-05-29 19:23:17 -06:00
POSTGRES_HOST: postgres
2019-11-27 20:57:21 -07:00
POSTGRES_DATABASE: dbname
POSTGRES_USER: user
POSTGRES_PASSWORD: password
```
2022-04-12 21:39:51 +03:00
2021-12-07 22:25:39 -05:00
- Images are tagged by the major PostgreSQL version they support: `10` , `11` , `12` , `13` , or `14` .
2021-12-07 22:15:47 -05:00
- The `SCHEDULE` variable determines backup frequency. See go-cron schedules documentation [here ](http://godoc.org/github.com/robfig/cron#hdr-Predefined_schedules ). Omit to run the backup immediately and then exit.
2019-12-01 18:47:18 -07:00
- If `PASSPHRASE` is provided, the backup will be encrypted using GPG.
2019-12-01 21:49:48 -07:00
- Run `docker exec <container name> sh backup.sh` to trigger a backup ad-hoc
2022-04-12 21:39:51 +03:00
- Use `BACKUP_KEEP_DAYS` to set time for how long you want to keep backup.
2019-11-27 20:57:21 -07:00
## Restore
2019-12-01 20:02:12 -07:00
> **WARNING:** DATA LOSS! All database objects will be dropped and re-created.
2019-12-01 18:47:18 -07:00
### ... from latest backup
2019-11-27 20:57:21 -07:00
```sh
docker exec < container name > sh restore.sh
```
2019-12-01 20:02:12 -07:00
> **NOTE:** If your bucket has more than a 1000 files, the latest may not be restored -- only one S3 `ls` command is used
2019-12-01 18:47:18 -07:00
### ... from specific backup
2019-11-29 22:17:47 -07:00
```sh
docker exec < container name > sh restore.sh < timestamp >
```
2019-12-01 21:49:48 -07:00
2022-07-03 12:23:22 -04:00
# Development
## Build the image locally
`ALPINE_VERSION` determines Postgres version compatibility. See [`build-and-push-images.yml` ](.github/workflows/build-and-push-images.yml ) for the latest mapping.
```sh
cd postgres-backup-s3
docker build --build-arg ALPINE_VERSION=3.14
```
2021-06-22 08:32:05 -07:00
# Acknowledgements
This project is a fork and re-structuring of @schickling 's [postgres-backup-s3 ](https://github.com/schickling/dockerfiles/tree/master/postgres-backup-s3 ) and [postgres-restore-s3 ](https://github.com/schickling/dockerfiles/tree/master/postgres-restore-s3 ).
2019-12-01 21:49:48 -07:00
## Fork goals
- [x] dedicated repository
- [x] automated builds
- [x] support multiple PostgreSQL versions
2020-05-29 19:23:17 -06:00
- [x] backup and restore with one image
2019-12-01 21:49:48 -07:00
- [x] support encrypted (password-protected) backups
- [x] option to restore from specific backup by timestamp
2022-04-13 00:34:08 +03:00
- [x] auto-removal old backups
2019-12-01 21:49:48 -07:00
## Other changes
- uses `pg_dump` 's `custom` format (see [docs ](https://www.postgresql.org/docs/10/app-pgdump.html ))
2020-05-28 23:21:48 -06:00
- doesn't use Python 2
2019-12-01 21:49:48 -07:00
- backup blobs and all schemas by default
- drop and re-create all database objects on restore
- some env vars renamed or removed
- filter backups on S3 by database name