update README, save some indentation

* adding option to skip tls verification error

* merge options

* removed merged option from README

Co-authored-by: Schwannden Kuo <schwannden@mobagel.com>

README.md

@@ -2,114 +2,48 @@
 
 Backup Docker volumes locally or to any S3 compatible storage.
 
-The [offen/docker-volume-backup](https://hub.docker.com/r/offen/docker-volume-backup) Docker image can be used as a lightweight (below 15MB) sidecar container to an existing Docker setup. It handles recurring backups of Docker volumes to a local directory or any S3 compatible storage (or both), and rotates away old backups if configured.
+The [offen/docker-volume-backup](https://hub.docker.com/r/offen/docker-volume-backup) Docker image can be used as a lightweight (below 15MB) sidecar container to an existing Docker setup.
+It handles __recurring or one-off backups of Docker volumes__ to a __local directory__ or __any S3 compatible storage__ (or both), and __rotates away old backups__ if configured. It also supports __encrypting your backups using GPG__ and __sending notifications for failed backup runs__.
 
-## Configuration
+<!-- MarkdownTOC -->
 
-Backup targets, schedule and retention are configured in environment variables:
+- [Quickstart](#quickstart)
+  - [Recurring backups in a compose setup](#recurring-backups-in-a-compose-setup)
+  - [One-off backups using Docker CLI](#one-off-backups-using-docker-cli)
+- [Configuration reference](#configuration-reference)
+- [How to](#how-to)
+  - [Stopping containers during backup](#stopping-containers-during-backup)
+  - [Automatically pruning old backups](#automatically-pruning-old-backups)
+  - [Send email notifications on failed backup runs](#send-email-notifications-on-failed-backup-runs)
+  - [Encrypting your backup using GPG](#encrypting-your-backup-using-gpg)
+  - [Restoring a volume from a backup](#restoring-a-volume-from-a-backup)
+  - [Set the timezone the container runs in](#set-the-timezone-the-container-runs-in)
+  - [Using with Docker Swarm](#using-with-docker-swarm)
+  - [Manually triggering a backup](#manually-triggering-a-backup)
+- [Recipes](#recipes)
+  - [Backing up to AWS S3](#backing-up-to-aws-s3)
+  - [Backing up to MinIO](#backing-up-to-minio)
+  - [Backing up locally](#backing-up-locally)
+  - [Backing up to AWS S3 as well as locally](#backing-up-to-aws-s3-as-well-as-locally)
+  - [Running on a custom cron schedule](#running-on-a-custom-cron-schedule)
+  - [Rotating away backups that are older than 7 days](#rotating-away-backups-that-are-older-than-7-days)
+  - [Encrypting your backups using GPG](#encrypting-your-backups-using-gpg)
+  - [Running multiple instances in the same setup](#running-multiple-instances-in-the-same-setup)
+- [Differences to `futurice/docker-volume-backup`](#differences-to-futuricedocker-volume-backup)
 
-```ini
-########### BACKUP SCHEDULE
+<!-- /MarkdownTOC -->
 
-# Backups run on the given cron schedule and use the filename defined in the
-# template expression.
+---
 
-BACKUP_CRON_EXPRESSION="0 2 * * *"
-# Format verbs will be replaced as in the `date` command. Omitting them
-# will result in the same filename for every backup run, which means previous
-# versions will be overwritten.
-BACKUP_FILENAME="backup-%Y-%m-%dT%H-%M-%S.tar.gz"
+Code and documentation for `v1` versions are found on [this branch][v1-branch].
 
-########### BACKUP STORAGE
+[v1-branch]: https://github.com/offen/docker-volume-backup/tree/v1
 
-# Define credentials for authenticating against the backup storage and a bucket
-# name. Although all of these values are `AWS`-prefixed, the setup can be used
-# with any S3 compatible storage.
+## Quickstart
 
-AWS_ACCESS_KEY_ID="<xxx>"
-AWS_SECRET_ACCESS_KEY="<xxx>"
-AWS_S3_BUCKET_NAME="<xxx>"
+### Recurring backups in a compose setup
 
-# This is the FQDN of your storage server, e.g. `storage.example.com`.
-# Do not set this when working against AWS S3. If you need to set a
-# specific protocol, you will need to use the option below.
-
-# AWS_ENDPOINT="<xxx>"
-
-# The protocol to be used when communicating with your storage server.
-# Defaults to "https". You can set this to "http" when communicating with
-# a different Docker container on the same host for example.
-
-# AWS_ENDPOINT_PROTO="https"
-
-# Setting this variable to any value will disable verification of
-# SSL certificates. You shouldn't use this unless you use self-signed
-# certificates for your remote storage backend.
-
-# AWS_ENDPOINT_INSECURE="true"
-
-# In addition to backing up you can also store backups locally. Pass in
-# a local path to store your backups here if needed. You likely want to
-# mount a local folder or Docker volume into that location when running
-# the container. Local paths can also be subject to pruning of old
-# backups as defined below.
-
-# BACKUP_ARCHIVE="/archive"
-
-########### BACKUP PRUNING
-
-# **IMPORTANT, PLEASE READ THIS BEFORE USING THIS FEATURE**:
-# The mechanism used for pruning backups is not very sophisticated
-# and applies its rules to **all files in the target directory** by default,
-# which means that if you are storing your backups next to other files,
-# these might become subject to deletion too. When using this option
-# make sure the backup files are stored in a directory used exclusively
-# for storing them or to configure BACKUP_PRUNING_PREFIX to limit
-# removal to certain files.
-
-# Define this value to enable automatic pruning of old backups. The value
-# declares the number of days for which a backup is kept.
-
-# BACKUP_RETENTION_DAYS="7"
-
-# In case the duration a backup takes fluctuates noticeably in your setup
-# you can adjust this setting to make sure there are no race conditions
-# between the backup finishing and the rotation not deleting backups that
-# sit on the very edge of the time window. Set this value to a duration
-# that is expected to be bigger than the maximum difference of backups.
-# Valid values have a suffix of (s)econds, (m)inutes or (h)ours.
-
-# BACKUP_PRUNING_LEEWAY="1m"
-
-# In case your target bucket or directory contains other files than the ones
-# managed by this container, you can limit the scope of rotation by setting
-# a prefix value. This would usually be the non-parametrized part of your
-# BACKUP_FILENAME. E.g. if BACKUP_FILENAME is `db-backup-%Y-%m-%dT%H-%M-%S.tar.gz`,
-# you can set BACKUP_PRUNING_PREFIX to `db-backup-` and make sure
-# unrelated files are not affected.
-
-# BACKUP_PRUNING_PREFIX="backup-"
-
-########### BACKUP ENCRYPTION
-
-# Backups can be encrypted using gpg in case a passphrase is given
-
-# GPG_PASSPHRASE="<xxx>"
-
-########### STOPPING CONTAINERS DURING BACKUP
-
-# Containers can be stopped by applying a
-# `docker-volume-backup.stop-during-backup` label. By default, all containers
-# that are labeled with `true` will be stopped. If you need more fine grained
-# control (e.g. when running multiple containers based on this image), you can
-# override this default by specifying a different value here.
-
-# BACKUP_STOP_CONTAINER_LABEL="service1"
-```
-
-## Example in a docker-compose setup
-
-Most likely, you will use this image as a sidecar container in an existing docker-compose setup like this:
+Add a `backup` service to your compose setup and mount the volumes you would like to see backed up:
 
 ```yml
 version: '3'
@@ -127,26 +61,331 @@ services:
       - docker-volume-backup.stop-during-backup=true
 
   backup:
+    # In production, it is advised to lock your image tag to a proper
+    # release version instead of using `latest`.
+    # Check https://github.com/offen/docker-volume-backup/releases
+    # for a list of available releases.
     image: offen/docker-volume-backup:latest
     restart: always
-    env_file: ./backup.env
+    env_file: ./backup.env # see below for configuration reference
     volumes:
+      - data:/backup/my-app-backup:ro
       # Mounting the Docker socket allows the script to stop and restart
       # the container during backup. You can omit this if you don't want
       # to stop the container
       - /var/run/docker.sock:/var/run/docker.sock:ro
-      - data:/backup/my-app-backup:ro
       # If you mount a local directory or volume to `/archive` a local
       # copy of the backup will be stored there. You can override the
-      # location inside of the container by setting `BACKUP_ARCHIVE`
-      # - /path/to/local_backups:/archive
+      # location inside of the container by setting `BACKUP_ARCHIVE`.
+      # You can omit this if you do not want to keep local backups.
+      - /path/to/local_backups:/archive
 volumes:
   data:
 ```
 
-## Using with Docker Swarm
+### One-off backups using Docker CLI
 
-By default, Docker Swarm will restart stopped containers automatically, even when manually stopped. If you plan to have your containers / services stopped during backup, this means you need to apply the `on-failure` restart policy to your service's definitions. A restart policy of `always` is not compatible with this tool.
+To run a one time backup, mount the volume you would like to see backed up into a container and run the `backup` command:
+
+```console
+docker run --rm \
+  -v data:/backup/data \
+  --env AWS_ACCESS_KEY_ID="<xxx>" \
+  --env AWS_SECRET_ACCESS_KEY="<xxx>" \
+  --env AWS_S3_BUCKET_NAME="<xxx>" \
+  --entrypoint backup \
+  offen/docker-volume-backup:latest
+```
+
+Alternatively, pass a `--env-file` in order to use a full config as described below.
+
+## Configuration reference
+
+Backup targets, schedule and retention are configured in environment variables.
+You can populate below template according to your requirements and use it as your `env_file`:
+
+```ini
+########### BACKUP SCHEDULE
+
+# Backups run on the given cron schedule in `busybox` flavor. If no
+# value is set, `@daily` will be used. If you do not want the cron
+# to ever run, use `0 0 5 31 2 ?`.
+
+# BACKUP_CRON_EXPRESSION="0 2 * * *"
+
+# The name of the backup file including the `.tar.gz` extension.
+# Format verbs will be replaced as in `strftime`. Omitting them
+# will result in the same filename for every backup run, which means previous
+# versions will be overwritten on subsequent runs. The default results
+# in filenames like `backup-2021-08-29T04-00-00.tar.gz`.
+
+# BACKUP_FILENAME="backup-%Y-%m-%dT%H-%M-%S.tar.gz"
+
+# When storing local backups, a symlink to the latest backup can be created
+# in case a value is given for this key. This has no effect on remote backups.
+
+# BACKUP_LATEST_SYMLINK="backup.latest.tar.gz"
+
+########### BACKUP STORAGE
+
+# The name of the remote bucket that should be used for storing backups. If
+# this is not set, no remote backups will be stored.
+
+# AWS_S3_BUCKET_NAME="backup-bucket"
+
+# Define credentials for authenticating against the backup storage and a bucket
+# name. Although all of these keys are `AWS`-prefixed, the setup can be used
+# with any S3 compatible storage.
+
+# AWS_ACCESS_KEY_ID="<xxx>"
+# AWS_SECRET_ACCESS_KEY="<xxx>"
+
+# Instead of providing static credentials, you can also use IAM instance profiles
+# or similar to provide authentication. Some possible configuration options on AWS:
+# - EC2: http://169.254.169.254
+# - ECS: http://169.254.170.2
+
+# AWS_IAM_ROLE_ENDPOINT="http://169.254.169.254"
+
+# This is the FQDN of your storage server, e.g. `storage.example.com`.
+# Do not set this when working against AWS S3 (the default value is 
+# `s3.amazonaws.com`). If you need to set a specific (non-https) protocol, you
+# will need to use the option below.
+
+# AWS_ENDPOINT="storage.example.com"
+
+# The protocol to be used when communicating with your storage server.
+# Defaults to "https". You can set this to "http" when communicating with
+# a different Docker container on the same host for example.
+
+# AWS_ENDPOINT_PROTO="https"
+
+# Setting this variable to `true` will disable verification of
+# SSL certificates. You shouldn't use this unless you use self-signed
+# certificates for your remote storage backend. This can only be used
+# when AWS_ENDPOINT_PROTO is set to `https`.
+
+# AWS_ENDPOINT_INSECURE="true"
+
+# In addition to storing backups remotely, you can also keep local copies.
+# Pass a container-local path to store your backups if needed. You also need to
+# mount a local folder or Docker volume into that location (`/archive`
+# by default) when running the container. In case the specified directory does
+# not exist (nothing is mounted) in the container when the backup is running,
+# local backups will be skipped. Local paths are also be subject to pruning of
+# old backups as defined below.
+
+# BACKUP_ARCHIVE="/archive"
+
+########### BACKUP PRUNING
+
+# **IMPORTANT, PLEASE READ THIS BEFORE USING THIS FEATURE**:
+# The mechanism used for pruning old backups is not very sophisticated
+# and applies its rules to **all files in the target directory** by default,
+# which means that if you are storing your backups next to other files,
+# these might become subject to deletion too. When using this option
+# make sure the backup files are stored in a directory used exclusively
+# for such files, or to configure BACKUP_PRUNING_PREFIX to limit
+# removal to certain files.
+
+# Define this value to enable automatic rotation of old backups. The value
+# declares the number of days for which a backup is kept.
+
+# BACKUP_RETENTION_DAYS="7"
+
+# In case the duration a backup takes fluctuates noticeably in your setup
+# you can adjust this setting to make sure there are no race conditions
+# between the backup finishing and the rotation not deleting backups that
+# sit on the edge of the time window. Set this value to a duration
+# that is expected to be bigger than the maximum difference of backups.
+# Valid values have a suffix of (s)econds, (m)inutes or (h)ours. By default,
+# one minute is used.
+
+# BACKUP_PRUNING_LEEWAY="1m"
+
+# In case your target bucket or directory contains other files than the ones
+# managed by this container, you can limit the scope of rotation by setting
+# a prefix value. This would usually be the non-parametrized part of your
+# BACKUP_FILENAME. E.g. if BACKUP_FILENAME is `db-backup-%Y-%m-%dT%H-%M-%S.tar.gz`,
+# you can set BACKUP_PRUNING_PREFIX to `db-backup-` and make sure
+# unrelated files are not affected by the rotation mechanism.
+
+# BACKUP_PRUNING_PREFIX="backup-"
+
+########### BACKUP ENCRYPTION
+
+# Backups can be encrypted using gpg in case a passphrase is given.
+
+# GPG_PASSPHRASE="<xxx>"
+
+########### STOPPING CONTAINERS DURING BACKUP
+
+# Containers can be stopped by applying a
+# `docker-volume-backup.stop-during-backup` label. By default, all containers
+# that are labeled with `true` will be stopped. If you need more fine grained
+# control (e.g. when running multiple containers based on this image), you can
+# override this default by specifying a different value here.
+
+# BACKUP_STOP_CONTAINER_LABEL="service1"
+
+########### EMAIL NOTIFICATIONS ON FAILED BACKUP RUNS
+
+# In case SMTP credentials are provided, notification emails can be sent out on
+# failed backup runs. These emails will contain the start time, the error
+# message and all log output prior to the failure.
+
+# The recipient(s) of the notification. Supply a comma separated list
+# of adresses if you want to notify multiple recipients. If this is
+# not set, no emails will be sent.
+
+# EMAIL_NOTIFICATION_RECIPIENT="you@example.com"
+
+# The "From" header of the sent email. Defaults to `noreply@nohost`.
+
+# EMAIL_NOTIFICATION_SENDER="no-reply@example.com"
+
+# Configuration and credentials for the SMTP server to be used.
+# EMAIL_SMTP_PORT defaults to 587.
+
+# EMAIL_SMTP_HOST="posteo.de"
+# EMAIL_SMTP_PASSWORD="<xxx>"
+# EMAIL_SMTP_USERNAME="no-reply@example.com"
+# EMAIL_SMTP_PORT="<port>"
+```
+
+## How to
+
+### Stopping containers during backup
+
+In many cases, it will be desirable to stop the services that are consuming the volume you want to backup in order to ensure data integrity.
+This image can automatically stop and restart containers and services (in case you are running Docker in Swarm mode).
+By default, any container that is labeled `docker-volume-backup.stop-during-backup=true` will be stopped before the backup is being taken and restarted once it has finished.
+
+In case you need more fine grained control about which containers should be stopped (e.g. when backing up multiple volumes on different schedules), you can set the `BACKUP_STOP_CONTAINER_LABEL` environment variable and then use the same value for labeling:
+
+```yml
+version: '3'
+
+services:
+  app:
+    # definition for app ...
+    labels:
+      - docker-volume-backup.stop-during-backup=service1
+
+  backup:
+    image: offen/docker-volume-backup:latest
+    environment:
+      BACKUP_STOP_CONTAINER_LABEL: service1
+    volumes:
+      - data:/backup/my-app-backup:ro
+      - /var/run/docker.sock:/var/run/docker.sock:ro
+
+volumes:
+  data:
+```
+
+### Automatically pruning old backups
+
+When `BACKUP_RETENTION_DAYS` is configured, the image will check if there are any backups in the remote bucket or local archive that are older than the given retention value and rotate these backups away.
+
+Be aware that this mechanism looks at __all files in the target bucket or archive__, which means that other files that are older than the given deadline are deleted as well. In case you need to use a target that cannot be used exclusively for your backups, you can configure `BACKUP_PRUNING_PREFIX` to limit which files are considered eligible for deletion:
+
+```yml
+version: '3'
+
+services:
+  # ... define other services using the `data` volume here
+  backup:
+    image: offen/docker-volume-backup:latest
+    environment:
+      BACKUP_FILENAME: backup-%Y-%m-%dT%H-%M-%S.tar.gz
+      BACKUP_PRUNING_PREFIX: backup-
+      BACKUP_RETENTION_DAYS: 7
+    volumes:
+      - ${HOME}/backups:/archive
+      - data:/backup/my-app-backup:ro
+      - /var/run/docker.sock:/var/run/docker.sock:ro
+
+volumes:
+  data:
+```
+
+### Send email notifications on failed backup runs
+
+To send out email notifications on failed backup runs, provide SMTP credentials, a sender and a recipient:
+
+```yml
+version: '3'
+
+services:
+  backup:
+    image: offen/docker-volume-backup:latest
+    environment:
+      # ... other configuration values go here
+      EMAIL_SMTP_HOST: "smtp.example.com"
+      EMAIL_SMTP_PASSWORD: "password"
+      EMAIL_SMTP_USERNAME: "username"
+      EMAIL_NOTIFICATION_SENDER: "noreply@example.com"
+      EMAIL_NOTIFICATION_RECIPIENT: "notifications@example.com"
+```
+
+### Encrypting your backup using GPG
+
+The image supports encrypting backups using GPG out of the box.
+In case a `GPG_PASSPHRASE` environment variable is set, the backup will be encrypted using the given key and saved as a `.gpg` file instead.
+
+Assuming you have `gpg` installed, you can decrypt such a backup using (your OS will prompt for the passphrase before decryption can happen):
+
+```console
+gpg -o backup.tar.gz -d backup.tar.gz.gpg
+```
+
+### Restoring a volume from a backup
+
+In case you need to restore a volume from a backup, the most straight forward procedure to do so would be:
+
+- Stop the container(s) that are using the volume
+- Untar the backup you want to restore
+  ```console
+  tar -C /tmp -xvf  backup.tar.gz
+  ```
+- Using a temporary one-off container, mount the volume (the example assumes it's named `data`) and copy over the backup. Make sure you copy the correct path level (this depends on how you mount your volume into the backup container), you might need to strip some leading elements
+  ```console
+  docker run -d --name backup_restore -v data:/backup_restore alpine
+  docker cp /tmp/backup/data-backup backup_restore:/backup_restore
+  docker stop backup_restore
+  docker rm backup_restore
+  ```
+- Restart the container(s) that are using the volume 
+
+Depending on your setup and the application(s) you are running, this might involve other steps to be taken still.
+
+### Set the timezone the container runs in
+
+By default a container based on this image will run in the UTC timezone.
+As the image is designed to be as small as possible, additional timezone data is not included.
+In case you want to run your cron rules in your local timezone (respecting DST and similar), you can mount your Docker host's `/etc/timezone` and `/etc/localtime` in read-only mode:
+
+```yml
+version: '3'
+
+services:
+  backup:
+    image: offen/docker-volume-backup:latest
+    volumes:
+      - data:/backup/my-app-backup:ro
+      - /etc/timezone:/etc/timezone:ro
+      - /etc/localtime:/etc/localtime:ro
+
+volumes:
+  data:
+```
+
+### Using with Docker Swarm
+
+By default, Docker Swarm will restart stopped containers automatically, even when manually stopped.
+If you plan to have your containers / services stopped during backup, this means you need to apply the `on-failure` restart policy to your service's definitions.
+A restart policy of `always` is not compatible with this tool.
 
 ---
 
@@ -162,7 +401,7 @@ services:
           memory: 25M
 ```
 
-## Manually triggering a backup
+### Manually triggering a backup
 
 You can manually trigger a backup run outside of the defined cron schedule by executing the `backup` command inside the container:
 
@@ -170,15 +409,212 @@ You can manually trigger a backup run outside of the defined cron schedule by ex
 docker exec <container_ref> backup
 ```
 
----
+## Recipes
+
+This section lists configuration for some real-world use cases that you can mix and match according to your needs.
+
+### Backing up to AWS S3
+
+```yml
+version: '3'
+
+services:
+  # ... define other services using the `data` volume here
+  backup:
+    image: offen/docker-volume-backup:latest
+    environment:
+      AWS_BUCKET_NAME: backup-bucket
+      AWS_ACCESS_KEY_ID: AKIAIOSFODNN7EXAMPLE
+      AWS_SECRET_ACCESS_KEY: wJalrXUtnFEMI/K7MDENG/bPxRfiCYEXAMPLEKEY
+    volumes:
+      - data:/backup/my-app-backup:ro
+      - /var/run/docker.sock:/var/run/docker.sock:ro
+
+volumes:
+  data:
+```
+
+### Backing up to MinIO
+
+```yml
+version: '3'
+
+services:
+  # ... define other services using the `data` volume here
+  backup:
+    image: offen/docker-volume-backup:latest
+    environment:
+      AWS_ENDPOINT: minio.example.com
+      AWS_BUCKET_NAME: backup-bucket
+      AWS_ACCESS_KEY_ID: MINIOACCESSKEY
+      AWS_SECRET_ACCESS_KEY: MINIOSECRETKEY
+    volumes:
+      - data:/backup/my-app-backup:ro
+      - /var/run/docker.sock:/var/run/docker.sock:ro
+
+volumes:
+  data:
+```
+
+### Backing up locally
+
+```yml
+version: '3'
+
+services:
+  # ... define other services using the `data` volume here
+  backup:
+    image: offen/docker-volume-backup:latest
+    environment:
+      BACKUP_FILENAME: backup-%Y-%m-%dT%H-%M-%S.tar.gz
+      BACKUP_LATEST_SYMLINK: backup-latest.tar.gz
+    volumes:
+      - data:/backup/my-app-backup:ro
+      - /var/run/docker.sock:/var/run/docker.sock:ro
+      - ${HOME}/backups:/archive
+
+volumes:
+  data:
+```
+
+### Backing up to AWS S3 as well as locally
+
+```yml
+version: '3'
+
+services:
+  # ... define other services using the `data` volume here
+  backup:
+    image: offen/docker-volume-backup:latest
+    environment:
+      AWS_BUCKET_NAME: backup-bucket
+      AWS_ACCESS_KEY_ID: AKIAIOSFODNN7EXAMPLE
+      AWS_SECRET_ACCESS_KEY: wJalrXUtnFEMI/K7MDENG/bPxRfiCYEXAMPLEKEY
+    volumes:
+      - data:/backup/my-app-backup:ro
+      - /var/run/docker.sock:/var/run/docker.sock:ro
+      - ${HOME}/backups:/archive
+
+volumes:
+  data:
+```
+
+### Running on a custom cron schedule
+
+```yml
+version: '3'
+
+services:
+  # ... define other services using the `data` volume here
+  backup:
+    image: offen/docker-volume-backup:latest
+    environment:
+      # take a backup on every hour
+      BACKUP_CRON_EXPRESSION: "0 * * * *"
+      AWS_BUCKET_NAME: backup-bucket
+      AWS_ACCESS_KEY_ID: AKIAIOSFODNN7EXAMPLE
+      AWS_SECRET_ACCESS_KEY: wJalrXUtnFEMI/K7MDENG/bPxRfiCYEXAMPLEKEY
+    volumes:
+      - data:/backup/my-app-backup:ro
+      - /var/run/docker.sock:/var/run/docker.sock:ro
+
+volumes:
+  data:
+```
+
+### Rotating away backups that are older than 7 days
+
+```yml
+version: '3'
+
+services:
+  # ... define other services using the `data` volume here
+  backup:
+    image: offen/docker-volume-backup:latest
+    environment:
+      AWS_BUCKET_NAME: backup-bucket
+      AWS_ACCESS_KEY_ID: AKIAIOSFODNN7EXAMPLE
+      AWS_SECRET_ACCESS_KEY: wJalrXUtnFEMI/K7MDENG/bPxRfiCYEXAMPLEKEY
+      BACKUP_FILENAME: backup-%Y-%m-%dT%H-%M-%S.tar.gz
+      BACKUP_PRUNING_PREFIX: backup-
+      BACKUP_RETENTION_DAYS: 7
+    volumes:
+      - data:/backup/my-app-backup:ro
+      - /var/run/docker.sock:/var/run/docker.sock:ro
+
+volumes:
+  data:
+```
+
+### Encrypting your backups using GPG
+
+```yml
+version: '3'
+
+services:
+  # ... define other services using the `data` volume here
+  backup:
+    image: offen/docker-volume-backup:latest
+    environment:
+      AWS_BUCKET_NAME: backup-bucket
+      AWS_ACCESS_KEY_ID: AKIAIOSFODNN7EXAMPLE
+      AWS_SECRET_ACCESS_KEY: wJalrXUtnFEMI/K7MDENG/bPxRfiCYEXAMPLEKEY
+      GPG_PASSPHRASE: somesecretstring
+    volumes:
+      - data:/backup/my-app-backup:ro
+      - /var/run/docker.sock:/var/run/docker.sock:ro
+
+volumes:
+  data:
+```
+
+### Running multiple instances in the same setup
+
+```yml
+version: '3'
+
+services:
+  # ... define other services using the `data_1` and `data_2` volumes here
+  backup_1: &backup_service
+    image: offen/docker-volume-backup:latest
+    environment: &backup_environment
+      BACKUP_CRON_EXPRESSION: "0 2 * * *"
+      AWS_BUCKET_NAME: backup-bucket
+      AWS_ACCESS_KEY_ID: AKIAIOSFODNN7EXAMPLE
+      AWS_SECRET_ACCESS_KEY: wJalrXUtnFEMI/K7MDENG/bPxRfiCYEXAMPLEKEY
+      # Label the container using the `data_1` volume as `docker-volume-backup.stop-during-backup=service1`
+      BACKUP_STOP_CONTAINER_LABEL: service1
+    volumes:
+      - data_1:/backup/data-1-backup:ro
+      - /var/run/docker.sock:/var/run/docker.sock:ro
+  backup_2:
+    <<: *backup_service
+    environment:
+      <<: *backup_environment
+      # Label the container using the `data_2` volume as `docker-volume-backup.stop-during-backup=service2`
+      BACKUP_CRON_EXPRESSION: "0 3 * * *"
+      BACKUP_STOP_CONTAINER_LABEL: service2
+    volumes:
+      - data_2:/backup/data-2-backup:ro
+      - /var/run/docker.sock:/var/run/docker.sock:ro
+
+volumes:
+  data_1:
+  data_2:
+```
 
 ## Differences to `futurice/docker-volume-backup`
 
-This image is heavily inspired by the `futurice/docker-volume-backup`. We decided to publish this image as a simpler and more lightweight alternative because of the following requirements:
+This image is heavily inspired by `futurice/docker-volume-backup`. We decided to publish this image as a simpler and more lightweight alternative because of the following requirements:
 
-- The original image is based on `ubuntu` and additional tools, making it very heavy. This version is roughly 1/25 in compressed size (it's ~12MB).
-- The original image uses a shell script, when this is written in Go, which makes it easier to extend and maintain (more verbose also).
-- The original image proposed to handle backup rotation through AWS S3 lifecycle policies. This image adds the option to rotate away old backups through the same command so this functionality can also be offered for non-AWS storage backends like MinIO. Local copies of backups can also be pruned once they reach a certain age.
+- The original image is based on `ubuntu` and requires additional tools, making it heavy.
+This version is roughly 1/25 in compressed size (it's ~12MB).
+- The original image uses a shell script, when this version is written in Go, which makes it easier to extend and maintain (more verbose too).
+- The original image proposed to handle backup rotation through AWS S3 lifecycle policies.
+This image adds the option to rotate away old backups through the same command so this functionality can also be offered for non-AWS storage backends like MinIO.
+Local copies of backups can also be pruned once they reach a certain age.
 - InfluxDB specific functionality from the original image was removed.
 - `arm64` and `arm/v7` architectures are supported.
 - Docker in Swarm mode is supported.
+- Notifications on failed backups are supported
+- IAM authentication through instance profiles is supported

cmd/backup/main.go

@@ -4,25 +4,29 @@
 package main
 
 import (
+	"bytes"
 	"context"
+	"errors"
 	"fmt"
 	"io"
 	"os"
 	"path"
 	"path/filepath"
+	"strings"
 	"time"
 
 	"github.com/docker/docker/api/types"
 	"github.com/docker/docker/api/types/filters"
 	"github.com/docker/docker/api/types/swarm"
 	"github.com/docker/docker/client"
+	"github.com/go-gomail/gomail"
 	"github.com/gofrs/flock"
 	"github.com/kelseyhightower/envconfig"
 	"github.com/leekchan/timeutil"
-	minio "github.com/minio/minio-go/v7"
+	"github.com/m90/targz"
+	"github.com/minio/minio-go/v7"
 	"github.com/minio/minio-go/v7/pkg/credentials"
 	"github.com/sirupsen/logrus"
-	"github.com/walle/targz"
 	"golang.org/x/crypto/openpgp"
 )
 
@@ -30,12 +34,39 @@ func main() {
 	unlock := lock("/var/lock/dockervolumebackup.lock")
 	defer unlock()
 
-	s := &script{}
-	s.must(s.init())
-	s.must(s.stopContainersAndRun(s.takeBackup))
-	s.must(s.encryptBackup())
-	s.must(s.copyBackup())
-	s.must(s.cleanBackup())
+	s, err := newScript()
+	if err != nil {
+		panic(err)
+	}
+
+	defer func() {
+		if err := recover(); err != nil {
+			if e, ok := err.(error); ok && strings.Contains(e.Error(), msgBackupFailed) {
+				os.Exit(1)
+			}
+			panic(err)
+		}
+	}()
+
+	s.must(func() error {
+		restartContainers, err := s.stopContainers()
+		defer func() {
+			s.must(restartContainers())
+		}()
+		if err != nil {
+			return err
+		}
+		return s.takeBackup()
+	}())
+
+	s.must(func() error {
+		defer func() {
+			s.must(s.removeArtifacts())
+		}()
+		s.must(s.encryptBackup())
+		return s.copyBackup()
+	}())
+
 	s.must(s.pruneOldBackups())
 	s.logger.Info("Finished running backup tasks.")
 }
@@ -43,99 +74,164 @@ func main() {
 // script holds all the stateful information required to orchestrate a
 // single backup run.
 type script struct {
-	ctx    context.Context
 	cli    *client.Client
 	mc     *minio.Client
 	logger *logrus.Logger
+	hooks  []hook
 
-	start time.Time
-	file  string
+	start  time.Time
+	file   string
+	output *bytes.Buffer
 
 	c *config
 }
 
 type config struct {
-	BackupSources            string        `split_words:"true" default:"/backup"`
-	BackupFilename           string        `split_words:"true" default:"backup-%Y-%m-%dT%H-%M-%S.tar.gz"`
-	BackupArchive            string        `split_words:"true" default:"/archive"`
-	BackupRetentionDays      int32         `split_words:"true" default:"-1"`
-	BackupPruningLeeway      time.Duration `split_words:"true" default:"1m"`
-	BackupPruningPrefix      string        `split_words:"true"`
-	BackupStopContainerLabel string        `split_words:"true" default:"true"`
-	AwsS3BucketName          string        `split_words:"true"`
-	AwsEndpoint              string        `split_words:"true" default:"s3.amazonaws.com"`
-	AwsEndpointProto         string        `split_words:"true" default:"https"`
-	AwsEndpointInsecure      bool          `split_words:"true"`
-	AwsAccessKeyID           string        `envconfig:"AWS_ACCESS_KEY_ID"`
-	AwsSecretAccessKey       string        `split_words:"true"`
-	GpgPassphrase            string        `split_words:"true"`
+	BackupSources              string        `split_words:"true" default:"/backup"`
+	BackupFilename             string        `split_words:"true" default:"backup-%Y-%m-%dT%H-%M-%S.tar.gz"`
+	BackupLatestSymlink        string        `split_words:"true"`
+	BackupArchive              string        `split_words:"true" default:"/archive"`
+	BackupRetentionDays        int32         `split_words:"true" default:"-1"`
+	BackupPruningLeeway        time.Duration `split_words:"true" default:"1m"`
+	BackupPruningPrefix        string        `split_words:"true"`
+	BackupStopContainerLabel   string        `split_words:"true" default:"true"`
+	AwsS3BucketName            string        `split_words:"true"`
+	AwsEndpoint                string        `split_words:"true" default:"s3.amazonaws.com"`
+	AwsEndpointProto           string        `split_words:"true" default:"https"`
+	AwsEndpointInsecure        bool          `split_words:"true"`
+	AwsAccessKeyID             string        `envconfig:"AWS_ACCESS_KEY_ID"`
+	AwsSecretAccessKey         string        `split_words:"true"`
+	AwsIamRoleEndpoint         string        `split_words:"true"`
+	GpgPassphrase              string        `split_words:"true"`
+	EmailNotificationRecipient string        `split_words:"true"`
+	EmailNotificationSender    string        `split_words:"true" default:"noreply@nohost"`
+	EmailSMTPHost              string        `envconfig:"EMAIL_SMTP_HOST"`
+	EmailSMTPPort              int           `envconfig:"EMAIL_SMTP_PORT" default:"587"`
+	EmailSMTPUsername          string        `envconfig:"EMAIL_SMTP_USERNAME"`
+	EmailSMTPPassword          string        `envconfig:"EMAIL_SMTP_PASSWORD"`
 }
 
-// init creates all resources needed for the script to perform actions against
+var msgBackupFailed = "backup run failed"
+
+// newScript creates all resources needed for the script to perform actions against
 // remote resources like the Docker engine or remote storage locations. All
 // reading from env vars or other configuration sources is expected to happen
 // in this method.
-func (s *script) init() error {
-	s.ctx = context.Background()
-	s.logger = logrus.New()
-	s.logger.SetOutput(os.Stdout)
-
-	s.c = &config{}
-	if err := envconfig.Process("", s.c); err != nil {
-		return fmt.Errorf("init: failed to process configuration values: %w", err)
+func newScript() (*script, error) {
+	stdOut, logBuffer := buffer(os.Stdout)
+	s := &script{
+		c: &config{},
+		logger: &logrus.Logger{
+			Out:       stdOut,
+			Formatter: new(logrus.TextFormatter),
+			Hooks:     make(logrus.LevelHooks),
+			Level:     logrus.InfoLevel,
+		},
+		start:  time.Now(),
+		output: logBuffer,
 	}
 
+	if err := envconfig.Process("", s.c); err != nil {
+		return nil, fmt.Errorf("newScript: failed to process configuration values: %w", err)
+	}
+
+	s.file = path.Join("/tmp", s.c.BackupFilename)
+
 	_, err := os.Stat("/var/run/docker.sock")
 	if !os.IsNotExist(err) {
 		cli, err := client.NewClientWithOpts(client.FromEnv, client.WithAPIVersionNegotiation())
 		if err != nil {
-			return fmt.Errorf("init: failed to create docker client")
+			return nil, fmt.Errorf("newScript: failed to create docker client")
 		}
 		s.cli = cli
 	}
 
 	if s.c.AwsS3BucketName != "" {
-		mc, err := minio.New(s.c.AwsEndpoint, &minio.Options{
-			Creds: credentials.NewStaticV4(
+		var creds *credentials.Credentials
+		if s.c.AwsAccessKeyID != "" && s.c.AwsSecretAccessKey != "" {
+			creds = credentials.NewStaticV4(
 				s.c.AwsAccessKeyID,
 				s.c.AwsSecretAccessKey,
 				"",
-			),
-			Secure: !s.c.AwsEndpointInsecure && s.c.AwsEndpointProto == "https",
-		})
+			)
+		} else if s.c.AwsIamRoleEndpoint != "" {
+			creds = credentials.NewIAM(s.c.AwsIamRoleEndpoint)
+		} else {
+			return nil, errors.New("newScript: AWS_S3_BUCKET_NAME is defined, but no credentials were provided")
+		}
+
+		options := minio.Options{
+			Creds:  creds,
+			Secure: s.c.AwsEndpointProto == "https",
+		}
+
+		if s.c.AwsEndpointInsecure {
+			if !options.Secure {
+				return nil, errors.New("newScript: AWS_ENDPOINT_INSECURE = true is only meaningful for https")
+			}
+
+			transport, err := minio.DefaultTransport(true)
+			if err != nil {
+				return nil, fmt.Errorf("newScript: failed to create default minio transport")
+			}
+			transport.TLSClientConfig.InsecureSkipVerify = true
+			options.Transport = transport
+		}
+
+		mc, err := minio.New(s.c.AwsEndpoint, &options)
 		if err != nil {
-			return fmt.Errorf("init: error setting up minio client: %w", err)
+			return nil, fmt.Errorf("newScript: error setting up minio client: %w", err)
 		}
 		s.mc = mc
 	}
 
-	s.file = path.Join("/tmp", s.c.BackupFilename)
-	s.start = time.Now()
+	if s.c.EmailNotificationRecipient != "" {
+		s.hooks = append(s.hooks, hook{hookLevelFailure, func(err error, start time.Time, logOutput string) error {
+			mailer := gomail.NewDialer(
+				s.c.EmailSMTPHost, s.c.EmailSMTPPort, s.c.EmailSMTPUsername, s.c.EmailSMTPPassword,
+			)
 
-	return nil
-}
+			subject := fmt.Sprintf(
+				"Failure running docker-volume-backup at %s", start.Format(time.RFC3339),
+			)
+			body := fmt.Sprintf(
+				"Running docker-volume-backup failed with error: %s\n\nLog output of the failed run was:\n\n%s\n", err, logOutput,
+			)
 
-// stopContainersAndRun stops all Docker containers that are marked as to being
-// stopped during the backup and runs the given thunk. After returning, it makes
-// sure containers are being restarted if required. In case the docker cli
-// is not configured, it will call the given thunk immediately.
-func (s *script) stopContainersAndRun(thunk func() error) error {
-	if s.cli == nil {
-		return thunk()
+			message := gomail.NewMessage()
+			message.SetHeader("From", s.c.EmailNotificationSender)
+			message.SetHeader("To", s.c.EmailNotificationRecipient)
+			message.SetHeader("Subject", subject)
+			message.SetBody("text/plain", body)
+			return mailer.DialAndSend(message)
+		}})
 	}
 
-	allContainers, err := s.cli.ContainerList(s.ctx, types.ContainerListOptions{
+	return s, nil
+}
+
+var noop = func() error { return nil }
+
+// stopContainers stops all Docker containers that are marked as to being
+// stopped during the backup and returns a function that can be called to
+// restart everything that has been stopped.
+func (s *script) stopContainers() (func() error, error) {
+	if s.cli == nil {
+		return noop, nil
+	}
+
+	allContainers, err := s.cli.ContainerList(context.Background(), types.ContainerListOptions{
 		Quiet: true,
 	})
 	if err != nil {
-		return fmt.Errorf("stopContainersAndRun: error querying for containers: %w", err)
+		return noop, fmt.Errorf("stopContainersAndRun: error querying for containers: %w", err)
 	}
 
 	containerLabel := fmt.Sprintf(
 		"docker-volume-backup.stop-during-backup=%s",
 		s.c.BackupStopContainerLabel,
 	)
-	containersToStop, err := s.cli.ContainerList(s.ctx, types.ContainerListOptions{
+	containersToStop, err := s.cli.ContainerList(context.Background(), types.ContainerListOptions{
 		Quiet: true,
 		Filters: filters.NewArgs(filters.KeyValuePair{
 			Key:   "label",
@@ -144,11 +240,11 @@ func (s *script) stopContainersAndRun(thunk func() error) error {
 	})
 
 	if err != nil {
-		return fmt.Errorf("stopContainersAndRun: error querying for containers to stop: %w", err)
+		return noop, fmt.Errorf("stopContainersAndRun: error querying for containers to stop: %w", err)
 	}
 
 	if len(containersToStop) == 0 {
-		return thunk()
+		return noop, nil
 	}
 
 	s.logger.Infof(
@@ -161,14 +257,23 @@ func (s *script) stopContainersAndRun(thunk func() error) error {
 	var stoppedContainers []types.Container
 	var stopErrors []error
 	for _, container := range containersToStop {
-		if err := s.cli.ContainerStop(s.ctx, container.ID, nil); err != nil {
+		if err := s.cli.ContainerStop(context.Background(), container.ID, nil); err != nil {
 			stopErrors = append(stopErrors, err)
 		} else {
 			stoppedContainers = append(stoppedContainers, container)
 		}
 	}
 
-	defer func() error {
+	var stopError error
+	if len(stopErrors) != 0 {
+		stopError = fmt.Errorf(
+			"stopContainersAndRun: %d error(s) stopping containers: %w",
+			len(stopErrors),
+			join(stopErrors...),
+		)
+	}
+
+	return func() error {
 		servicesRequiringUpdate := map[string]struct{}{}
 
 		var restartErrors []error
@@ -177,13 +282,13 @@ func (s *script) stopContainersAndRun(thunk func() error) error {
 				servicesRequiringUpdate[swarmServiceName] = struct{}{}
 				continue
 			}
-			if err := s.cli.ContainerStart(s.ctx, container.ID, types.ContainerStartOptions{}); err != nil {
+			if err := s.cli.ContainerStart(context.Background(), container.ID, types.ContainerStartOptions{}); err != nil {
 				restartErrors = append(restartErrors, err)
 			}
 		}
 
 		if len(servicesRequiringUpdate) != 0 {
-			services, _ := s.cli.ServiceList(s.ctx, types.ServiceListOptions{})
+			services, _ := s.cli.ServiceList(context.Background(), types.ServiceListOptions{})
 			for serviceName := range servicesRequiringUpdate {
 				var serviceMatch swarm.Service
 				for _, service := range services {
@@ -193,11 +298,11 @@ func (s *script) stopContainersAndRun(thunk func() error) error {
 					}
 				}
 				if serviceMatch.ID == "" {
-					return fmt.Errorf("stopContainersAndRun: Couldn't find service with name %s", serviceName)
+					return fmt.Errorf("stopContainersAndRun: couldn't find service with name %s", serviceName)
 				}
 				serviceMatch.Spec.TaskTemplate.ForceUpdate = 1
 				_, err := s.cli.ServiceUpdate(
-					s.ctx, serviceMatch.ID,
+					context.Background(), serviceMatch.ID,
 					serviceMatch.Version, serviceMatch.Spec, types.ServiceUpdateOptions{},
 				)
 				if err != nil {
@@ -210,7 +315,7 @@ func (s *script) stopContainersAndRun(thunk func() error) error {
 			return fmt.Errorf(
 				"stopContainersAndRun: %d error(s) restarting containers and services: %w",
 				len(restartErrors),
-				err,
+				join(restartErrors...),
 			)
 		}
 		s.logger.Infof(
@@ -218,17 +323,7 @@ func (s *script) stopContainersAndRun(thunk func() error) error {
 			len(stoppedContainers),
 		)
 		return nil
-	}()
-
-	if len(stopErrors) != 0 {
-		return fmt.Errorf(
-			"stopContainersAndRun: %d error(s) stopping containers: %w",
-			len(stopErrors),
-			err,
-		)
-	}
-
-	return thunk()
+	}, stopError
 }
 
 // takeBackup creates a tar archive of the configured backup location and
@@ -244,11 +339,12 @@ func (s *script) takeBackup() error {
 
 // encryptBackup encrypts the backup file using PGP and the configured passphrase.
 // In case no passphrase is given it returns early, leaving the backup file
-//  untouched.
+// untouched.
 func (s *script) encryptBackup() error {
 	if s.c.GpgPassphrase == "" {
 		return nil
 	}
+	defer os.Remove(s.file)
 
 	gpgFile := fmt.Sprintf("%s.gpg", s.file)
 	outFile, err := os.Create(gpgFile)
@@ -276,10 +372,6 @@ func (s *script) encryptBackup() error {
 		return fmt.Errorf("encryptBackup: error writing ciphertext to file: %w", err)
 	}
 
-	if err := os.Remove(s.file); err != nil {
-		return fmt.Errorf("encryptBackup: error removing unencrpyted backup: %w", err)
-	}
-
 	s.file = gpgFile
 	s.logger.Infof("Encrypted backup using given passphrase, saving as `%s`.", s.file)
 	return nil
@@ -289,31 +381,48 @@ func (s *script) encryptBackup() error {
 // as per the given configuration.
 func (s *script) copyBackup() error {
 	_, name := path.Split(s.file)
-	if s.c.AwsS3BucketName != "" {
-		_, err := s.mc.FPutObject(s.ctx, s.c.AwsS3BucketName, name, s.file, minio.PutObjectOptions{
+	if s.mc != nil {
+		_, err := s.mc.FPutObject(context.Background(), s.c.AwsS3BucketName, name, s.file, minio.PutObjectOptions{
 			ContentType: "application/tar+gzip",
 		})
 		if err != nil {
 			return fmt.Errorf("copyBackup: error uploading backup to remote storage: %w", err)
 		}
-		s.logger.Infof("Uploaded a copy of backup `%s` to bucket `%s`", s.file, s.c.AwsS3BucketName)
+		s.logger.Infof("Uploaded a copy of backup `%s` to bucket `%s`.", s.file, s.c.AwsS3BucketName)
 	}
 
 	if _, err := os.Stat(s.c.BackupArchive); !os.IsNotExist(err) {
 		if err := copy(s.file, path.Join(s.c.BackupArchive, name)); err != nil {
 			return fmt.Errorf("copyBackup: error copying file to local archive: %w", err)
 		}
-		s.logger.Infof("Stored copy of backup `%s` in local archive `%s`", s.file, s.c.AwsS3BucketName)
+		s.logger.Infof("Stored copy of backup `%s` in local archive `%s`.", s.file, s.c.BackupArchive)
+		if s.c.BackupLatestSymlink != "" {
+			symlink := path.Join(s.c.BackupArchive, s.c.BackupLatestSymlink)
+			if _, err := os.Lstat(symlink); err == nil {
+				os.Remove(symlink)
+			}
+			if err := os.Symlink(name, symlink); err != nil {
+				return fmt.Errorf("copyBackup: error creating latest symlink: %w", err)
+			}
+			s.logger.Infof("Created/Updated symlink `%s` for latest backup.", s.c.BackupLatestSymlink)
+		}
 	}
 	return nil
 }
 
-// cleanBackup removes the backup file from disk.
-func (s *script) cleanBackup() error {
-	if err := os.Remove(s.file); err != nil {
-		return fmt.Errorf("cleanBackup: error removing file: %w", err)
+// removeArtifacts removes the backup file from disk.
+func (s *script) removeArtifacts() error {
+	_, err := os.Stat(s.file)
+	if err != nil {
+		if os.IsNotExist(err) {
+			return nil
+		}
+		return fmt.Errorf("removeArtifacts: error calling stat on file %s: %w", s.file, err)
 	}
-	s.logger.Info("Cleaned up local artifacts.")
+	if err := os.Remove(s.file); err != nil {
+		return fmt.Errorf("removeArtifacts: error removing file %s: %w", s.file, err)
+	}
+	s.logger.Infof("Removed local artifacts %s.", s.file)
 	return nil
 }
 
@@ -330,11 +439,10 @@ func (s *script) pruneOldBackups() error {
 		time.Sleep(s.c.BackupPruningLeeway)
 	}
 
-	s.logger.Infof("Trying to prune backups older than %d day(s) now.", s.c.BackupRetentionDays)
 	deadline := time.Now().AddDate(0, 0, -int(s.c.BackupRetentionDays))
 
-	if s.c.AwsS3BucketName != "" {
-		candidates := s.mc.ListObjects(s.ctx, s.c.AwsS3BucketName, minio.ListObjectsOptions{
+	if s.mc != nil {
+		candidates := s.mc.ListObjects(context.Background(), s.c.AwsS3BucketName, minio.ListObjectsOptions{
 			WithMetadata: true,
 			Prefix:       s.c.BackupPruningPrefix,
 		})
@@ -362,25 +470,26 @@ func (s *script) pruneOldBackups() error {
 				}
 				close(objectsCh)
 			}()
-			errChan := s.mc.RemoveObjects(s.ctx, s.c.AwsS3BucketName, objectsCh, minio.RemoveObjectsOptions{})
-			var errors []error
+			errChan := s.mc.RemoveObjects(context.Background(), s.c.AwsS3BucketName, objectsCh, minio.RemoveObjectsOptions{})
+			var removeErrors []error
 			for result := range errChan {
 				if result.Err != nil {
-					errors = append(errors, result.Err)
+					removeErrors = append(removeErrors, result.Err)
 				}
 			}
 
-			if len(errors) != 0 {
+			if len(removeErrors) != 0 {
 				return fmt.Errorf(
 					"pruneOldBackups: %d error(s) removing files from remote storage: %w",
-					len(errors),
-					errors[0],
+					len(removeErrors),
+					join(removeErrors...),
 				)
 			}
 			s.logger.Infof(
-				"Pruned %d out of %d remote backup(s) as their age exceeded the configured retention period.",
+				"Pruned %d out of %d remote backup(s) as their age exceeded the configured retention period of %d days.",
 				len(matches),
 				lenCandidates,
+				s.c.BackupRetentionDays,
 			)
 		} else if len(matches) != 0 && len(matches) == lenCandidates {
 			s.logger.Warnf(
@@ -414,29 +523,30 @@ func (s *script) pruneOldBackups() error {
 				)
 			}
 
-			if fi.ModTime().Before(deadline) {
+			if fi.Mode() != os.ModeSymlink && fi.ModTime().Before(deadline) {
 				matches = append(matches, candidate)
 			}
 		}
 
 		if len(matches) != 0 && len(matches) != len(candidates) {
-			var errors []error
+			var removeErrors []error
 			for _, candidate := range matches {
 				if err := os.Remove(candidate); err != nil {
-					errors = append(errors, err)
+					removeErrors = append(removeErrors, err)
 				}
 			}
-			if len(errors) != 0 {
+			if len(removeErrors) != 0 {
 				return fmt.Errorf(
 					"pruneOldBackups: %d error(s) deleting local files, starting with: %w",
-					len(errors),
-					errors[0],
+					len(removeErrors),
+					join(removeErrors...),
 				)
 			}
 			s.logger.Infof(
-				"Pruned %d out of %d local backup(s) as their age exceeded the configured retention period.",
+				"Pruned %d out of %d local backup(s) as their age exceeded the configured retention period of %d days.",
 				len(matches),
 				len(candidates),
+				s.c.BackupRetentionDays,
 			)
 		} else if len(matches) != 0 && len(matches) == len(candidates) {
 			s.logger.Warnf(
@@ -451,13 +561,35 @@ func (s *script) pruneOldBackups() error {
 	return nil
 }
 
+// runHooks runs all hooks that have been registered using the
+// given level. In case executing a hook returns an error, the following
+// hooks will still be run before the function returns an error.
+func (s *script) runHooks(err error, targetLevel string) error {
+	var actionErrors []error
+	for _, hook := range s.hooks {
+		if hook.level != targetLevel {
+			continue
+		}
+		if err := hook.action(err, s.start, s.output.String()); err != nil {
+			actionErrors = append(actionErrors, err)
+		}
+	}
+	if len(actionErrors) != 0 {
+		return join(actionErrors...)
+	}
+	return nil
+}
+
+// must exits the script run prematurely in case the given error
+// is non-nil. If failure hooks have been registered on the script object, they
+// will be called, passing the failure and previous log output.
 func (s *script) must(err error) {
 	if err != nil {
-		if s.logger == nil {
-			panic(err)
-		}
 		s.logger.Errorf("Fatal error running backup: %s", err)
-		os.Exit(1)
+		if hookErr := s.runHooks(err, hookLevelFailure); hookErr != nil {
+			s.logger.Errorf("An error occurred calling the registered failure hooks: %s", hookErr)
+		}
+		panic(errors.New(msgBackupFailed))
 	}
 }
 
@@ -496,3 +628,48 @@ func copy(src, dst string) error {
 	}
 	return out.Close()
 }
+
+// join takes a list of errors and joins them into a single error
+func join(errs ...error) error {
+	if len(errs) == 1 {
+		return errs[0]
+	}
+	var msgs []string
+	for _, err := range errs {
+		if err == nil {
+			continue
+		}
+		msgs = append(msgs, err.Error())
+	}
+	return errors.New("[" + strings.Join(msgs, ", ") + "]")
+}
+
+// buffer takes an io.Writer and returns a wrapped version of the
+// writer that writes to both the original target as well as the returned buffer
+func buffer(w io.Writer) (io.Writer, *bytes.Buffer) {
+	buffering := &bufferingWriter{buf: bytes.Buffer{}, writer: w}
+	return buffering, &buffering.buf
+}
+
+type bufferingWriter struct {
+	buf    bytes.Buffer
+	writer io.Writer
+}
+
+func (b *bufferingWriter) Write(p []byte) (n int, err error) {
+	if n, err := b.buf.Write(p); err != nil {
+		return n, fmt.Errorf("bufferingWriter: error writing to buffer: %w", err)
+	}
+	return b.writer.Write(p)
+}
+
+// hook contains a queued action that can be trigger them when the script
+// reaches a certain point (e.g. unsuccessful backup)
+type hook struct {
+	level  string
+	action func(err error, start time.Time, logOutput string) error
+}
+
+const (
+	hookLevelFailure = "failure"
+)

go.mod

@@ -7,9 +7,9 @@ require (
 	github.com/gofrs/flock v0.8.1
 	github.com/kelseyhightower/envconfig v1.4.0
 	github.com/leekchan/timeutil v0.0.0-20150802142658-28917288c48d
+	github.com/m90/targz v0.0.0-20210904082215-2e9a4529a615
 	github.com/minio/minio-go/v7 v7.0.12
 	github.com/sirupsen/logrus v1.8.1
-	github.com/walle/targz v0.0.0-20140417120357-57fe4206da5a
 	golang.org/x/crypto v0.0.0-20210817164053-32db794688a5
 )
 
@@ -20,6 +20,7 @@ require (
 	github.com/docker/go-connections v0.4.0 // indirect
 	github.com/docker/go-units v0.4.0 // indirect
 	github.com/dustin/go-humanize v1.0.0 // indirect
+	github.com/go-gomail/gomail v0.0.0-20160411212932-81ebce5c23df // indirect
 	github.com/gogo/protobuf v1.3.2 // indirect
 	github.com/golang/protobuf v1.5.0 // indirect
 	github.com/google/uuid v1.2.0 // indirect
@@ -41,5 +42,6 @@ require (
 	google.golang.org/genproto v0.0.0-20201110150050-8816d57aaa9a // indirect
 	google.golang.org/grpc v1.33.2 // indirect
 	google.golang.org/protobuf v1.26.0 // indirect
+	gopkg.in/alexcesaro/quotedprintable.v3 v3.0.0-20150716171945-2caba252f4dc // indirect
 	gopkg.in/ini.v1 v1.57.0 // indirect
 )

go.sum

@@ -254,6 +254,8 @@ github.com/ghodss/yaml v1.0.0/go.mod h1:4dBDuWmgqj2HViK6kFavaiC9ZROes6MMH2rRYeME
 github.com/go-gl/glfw v0.0.0-20190409004039-e6da0acd62b1/go.mod h1:vR7hzQXu2zJy9AVAgeJqvqgH9Q5CA+iKCZ2gyEVpxRU=
 github.com/go-gl/glfw/v3.3/glfw v0.0.0-20191125211704-12ad95a8df72/go.mod h1:tQ2UAYgL5IevRw8kRxooKSPJfGvJ9fJQFa0TUsXzTg8=
 github.com/go-gl/glfw/v3.3/glfw v0.0.0-20200222043503-6f7a984d4dc4/go.mod h1:tQ2UAYgL5IevRw8kRxooKSPJfGvJ9fJQFa0TUsXzTg8=
+github.com/go-gomail/gomail v0.0.0-20160411212932-81ebce5c23df h1:Bao6dhmbTA1KFVxmJ6nBoMuOJit2yjEgLJpIMYpop0E=
+github.com/go-gomail/gomail v0.0.0-20160411212932-81ebce5c23df/go.mod h1:GJr+FCSXshIwgHBtLglIg9M2l2kQSi6QjVAngtzI08Y=
 github.com/go-ini/ini v1.25.4/go.mod h1:ByCAeIL28uOIIG0E3PJtZPDL8WnHpFKFOtgjp+3Ies8=
 github.com/go-kit/kit v0.8.0/go.mod h1:xBxKIO96dXMWWy0MnWVtmwkA9/13aqxPnvrjFYMA2as=
 github.com/go-kit/kit v0.9.0/go.mod h1:xBxKIO96dXMWWy0MnWVtmwkA9/13aqxPnvrjFYMA2as=
@@ -407,6 +409,8 @@ github.com/kr/text v0.1.0 h1:45sCR5RtlFHMR4UwH9sdQ5TC8v0qDQCHnXt+kaKSTVE=
 github.com/kr/text v0.1.0/go.mod h1:4Jbv+DJW3UT/LiOwJeYQe1efqtUx/iVham/4vfdArNI=
 github.com/leekchan/timeutil v0.0.0-20150802142658-28917288c48d h1:2puqoOQwi3Ai1oznMOsFIbifm6kIfJaLLyYzWD4IzTs=
 github.com/leekchan/timeutil v0.0.0-20150802142658-28917288c48d/go.mod h1:hO90vCP2x3exaSH58BIAowSKvV+0OsY21TtzuFGHON4=
+github.com/m90/targz v0.0.0-20210904082215-2e9a4529a615 h1:rn0LO2tQEgCDOct8qnbcslTUpAIWdVlWcGkjoumhf2U=
+github.com/m90/targz v0.0.0-20210904082215-2e9a4529a615/go.mod h1:YZK3bSO/oVlk9G+v00BxgzxW2Us4p/R4ysHOBjk0fJI=
 github.com/magiconair/properties v1.8.0/go.mod h1:PppfXfuXeibc/6YijjN8zIbojt8czPbwD3XqdrwzmxQ=
 github.com/mailru/easyjson v0.0.0-20190614124828-94de47d64c63/go.mod h1:C1wdFJiN94OJF2b5HbByQZoLdCWB1Yqtg26g4irojpc=
 github.com/mailru/easyjson v0.0.0-20190626092158-b2ccc519800e/go.mod h1:C1wdFJiN94OJF2b5HbByQZoLdCWB1Yqtg26g4irojpc=
@@ -596,8 +600,6 @@ github.com/vishvananda/netlink v1.1.1-0.20201029203352-d40f9887b852/go.mod h1:tw
 github.com/vishvananda/netns v0.0.0-20180720170159-13995c7128cc/go.mod h1:ZjcWmFBXmLKZu9Nxj3WKYEafiSqer2rnvPr0en9UNpI=
 github.com/vishvananda/netns v0.0.0-20191106174202-0a2b9b5464df/go.mod h1:JP3t17pCcGlemwknint6hfoeCVQrEMVwxRLRjXpq+BU=
 github.com/vishvananda/netns v0.0.0-20200728191858-db3c7e526aae/go.mod h1:DD4vA1DwXk04H54A1oHXtwZmA0grkVMdPxx/VGLCah0=
-github.com/walle/targz v0.0.0-20140417120357-57fe4206da5a h1:6cKSHLRphD9Fo1LJlISiulvgYCIafJ3QfKLimPYcAGc=
-github.com/walle/targz v0.0.0-20140417120357-57fe4206da5a/go.mod h1:nccQrXCnc5SjsThFLmL7hYbtT/mHJcuolPifzY5vJqE=
 github.com/willf/bitset v1.1.11-0.20200630133818-d5bec3311243/go.mod h1:RjeCKbqT1RxIR/KWY6phxZiaY1IyutSBfGjNPySAYV4=
 github.com/willf/bitset v1.1.11/go.mod h1:83CECat5yLh5zVOf4P1ErAgKA5UDvKtgyUABdr3+MjI=
 github.com/xeipuuv/gojsonpointer v0.0.0-20180127040702-4e3ac2762d5f/go.mod h1:N2zxlSyiKSe5eX1tZViRH5QA0qijqEDrYZiPEAiq3wU=
@@ -908,6 +910,8 @@ google.golang.org/protobuf v1.26.0 h1:bxAC2xTBsZGibn2RTntX0oH50xLsqy1OxA9tTL3p/l
 google.golang.org/protobuf v1.26.0/go.mod h1:9q0QmTI4eRPtz6boOQmLYwt+qCgq0jsYwAQnmE0givc=
 gopkg.in/airbrake/gobrake.v2 v2.0.9/go.mod h1:/h5ZAUhDkGaJfjzjKLSjv6zCL6O0LLBxU4K+aSYdM/U=
 gopkg.in/alecthomas/kingpin.v2 v2.2.6/go.mod h1:FMv+mEhP44yOT+4EoQTLFTRgOQ1FBLkstjWtayDeSgw=
+gopkg.in/alexcesaro/quotedprintable.v3 v3.0.0-20150716171945-2caba252f4dc h1:2gGKlE2+asNV9m7xrywl36YYNnBG5ZQ0r/BOOxqPpmk=
+gopkg.in/alexcesaro/quotedprintable.v3 v3.0.0-20150716171945-2caba252f4dc/go.mod h1:m7x9LTH6d71AHyAX77c9yqWCCa3UKHcVEj9y7hAtKDk=
 gopkg.in/check.v1 v0.0.0-20161208181325-20d25e280405/go.mod h1:Co6ibVJAznAaIkqp8huTwlJQCZ016jof/cbN4VW5Yz0=
 gopkg.in/check.v1 v1.0.0-20141024133853-64131543e789/go.mod h1:Co6ibVJAznAaIkqp8huTwlJQCZ016jof/cbN4VW5Yz0=
 gopkg.in/check.v1 v1.0.0-20180628173108-788fd7840127/go.mod h1:Co6ibVJAznAaIkqp8huTwlJQCZ016jof/cbN4VW5Yz0=

test/cli/run.sh

@@ -29,8 +29,7 @@ docker run -d \
 
 sleep 10
 
-docker run -d \
-  --name backup \
+docker run --rm \
   --network test_network \
   -v app_data:/backup/app_data \
   -v /var/run/docker.sock:/var/run/docker.sock \
@@ -40,18 +39,16 @@ docker run -d \
   --env AWS_ENDPOINT_PROTO=http \
   --env AWS_S3_BUCKET_NAME=backup \
   --env BACKUP_FILENAME=test.tar.gz \
-  --env BACKUP_CRON_EXPRESSION="0 0 5 31 2 ?" \
+  --entrypoint backup \
   offen/docker-volume-backup:$TEST_VERSION
 
-docker exec backup backup
-
 docker run --rm -it \
   -v backup_data:/data alpine \
   ash -c 'tar -xvf /data/backup/test.tar.gz && test -f /backup/app_data/offen.db'
 
 echo "[TEST:PASS] Found relevant files in untared backup."
 
-if [ "$(docker ps -q | wc -l)" != "3" ]; then
+if [ "$(docker ps -q | wc -l)" != "2" ]; then
   echo "[TEST:FAIL] Expected all containers to be running post backup, instead seen:"
   docker ps
   exit 1

test/compose/docker-compose.yml

@@ -24,6 +24,7 @@ services:
       AWS_ENDPOINT_PROTO: http
       AWS_S3_BUCKET_NAME: backup
       BACKUP_FILENAME: test.tar.gz
+      BACKUP_LATEST_SYMLINK: test.latest.tar.gz.gpg
       BACKUP_CRON_EXPRESSION: 0 0 5 31 2 ?
       BACKUP_RETENTION_DAYS: ${BACKUP_RETENTION_DAYS:-7}
       BACKUP_PRUNING_LEEWAY: 5s

test/compose/run.sh

@@ -9,6 +9,7 @@ mkdir -p local
 docker-compose up -d
 sleep 5
 
+docker-compose exec offen ln -s /var/opt/offen/offen.db /var/opt/offen/db.link
 docker-compose exec backup backup
 
 docker run --rm -it \
@@ -17,9 +18,11 @@ docker run --rm -it \
 
 echo "[TEST:PASS] Found relevant files in untared remote backup."
 
+test -L ./local/test.latest.tar.gz.gpg
 echo 1234secret | gpg -d --yes --passphrase-fd 0 ./local/test.tar.gz.gpg > ./local/decrypted.tar.gz
 tar -xf ./local/decrypted.tar.gz -C /tmp && test -f /tmp/backup/app_data/offen.db
 rm ./local/decrypted.tar.gz
+test -L /tmp/backup/app_data/db.link
 
 echo "[TEST:PASS] Found relevant files in untared local backup."