This container contains a shell script to dump and restore all databases of an ArangoDB server (Single Server or Cluster).
You must be able to reach the server from inside the running container. Please refer to the orchestration guide of your plattform.
It is important to use a user that has access to all database you want to dump.
There are two ways to authenticaton in the cluster: using a username/password (ARANGO_USERNAME and ARANGO_PASSWD) or a JWT secret (not yet supported).
In a Kubernetes environment you should store either the password or the JWT secret in a Kubernetes secret and pass it into the pod. If you are using the Kubernetes operator, then the JWT secret will already be available as Kubernetes secret.
For testing purposes you can obviously also pass the password or token directly in the environment. However, you should avoid this setup in a production environment.
For other setups, you can also use ARANGODB_PASSWD_FILE to pass the password as a file instead of an environment variable. Please not that this file is taken without any modification. You need to ensure that it does not contain any superfluous newlines or carriage returns.
usage: backup.sh (dump|restore)
the following environment variables will be recognized:
- ARANGO_ENDPOINT
- ARANGO_USERNAME
- ARANGO_PASSWD
- ARANGO_CLEAN_DUMP_DIRECTORY
- ARANGO_DUMP_DIRECTORY
- ARANGO_DUMP_THREADS
- ARANGO_LOGFILE
The default dump directory is /backup/dump. If you run the docker
container then the /backup dictory should be mounted to a volume
large enough to hold the dump data.
A typical usage is
docker run -v /some/large/disk:/backup -e ARANGO_ENDPOINT=tcp://server.name:8529 -e ARANGO_CLEAN_DUMP_DIRECTORY=true arangodb/arangodb-backup backup
dump uses arangodump and restore uses arangodump underneath. See Backup & Restore
for details about these programs.