Expand description
§Manage S3 and S3 backups
Used to manage objects created by Tocco.
See also S3 Storage Design Overview.
§Code Overview
| Where | What |
|---|---|
fs::Archive | Representationof a backup archive and its configuration. |
state::Db | State DB for tracking live objects. |
db | Obtaining list of objects from DB servers. |
§Configuration
Archives are configured via config.toml in the archive’s root directory. Run
tocco-s3 init to initialize a repository with a default config, then modify
it as needed:
[archive]
# Minimum number of object that must exist in archive.
#
# (optional, default = 1)
# min_objects = 1
# Path to state DB.
#
# (optional, default="state.sqlite")
#
# Relitive paths are relative to the directory containing this config.
# state_db_path = "state.sqlite"
[s3]
endpoint = "https://objects.rma.cloudscale.ch"
# Either set an API key for Cloudscale …
cloudscale_api_key = "<unset>"
# … or list all buckets with keys.
#
# During regular operations cloudscale_api_key should be used. This option exists
# to ease restoring on a non-Cloudscale S3 cluster in case of an emergency.
#
# [s3.keys.bucket_name]
# access_key = "<unset>"
# secret_key = "<unset>"
[removal]
# Objects that have not been reference by Nice database
# (prod or test) are removed.
#
# (min=90)
retention_time_days = 180
# Enable dry run which only prints what be deleted
#
# (optional, default: false)
dry_run = true
[[db_servers]]
host_name = "db.example.net"
# (optional)
ssh_user = "postgres"
# (optional)
ssh_port = 22
# Minimum number of objects that must exist on server. Enforced
# by `tocco-s3 check-availability`.
#
# (optional, default=1)
# min_objects = 1