2.0.1

Release 2.0.1

2.0.0

Changelog

The software has been almost entirely rewritten.

Here the new features:

• Bivac now runs as a daemon, refresh the volume list every 10 minutes and backups volumes every day.
• The support of multiple backup engines has been dropped to focus on the Restic integration.
• A client is available and lets you manage your backups. You can manually force a backup, watch the logs or run custom Restic commands.
• An API is available to lets you build tools around Bivac.
• The API exposes a Prometheus endpoint from where you can collect metrics about Bivac and the backups.
• The scheduler can run multiple backups at the same time (at most 2 by host).

1.0.0-alpha8

Changelog

BREAKING CHANGES:

• core: Remove extra-env arguments (all manager environments are now forwarded to worker)
• core: Remove old AWS and Swift arguments. Please use environment variable forwarding introduced in alpha5

IMPROVEMENTS:

• core: Allow to configure engines commands
• orechestrator/docker: forward manager volumes to worker

BUG FIXES:

• orchestrator/kubernetes: Use namespace instead of hostname as instance label for metrics
• engine/restic: Use --host instead of deprecated --hostname by default
• engine/restic: Fix hostname in Kubernetes so that forget works properly
• providers: Ignore provider detection failure when container has no shell

1.0.0-alpha7

Changelog

IMPROVEMENTS:

• orchestrator/cattle: improve error handling
• core: improve template for Openshift to allow usage of custom providers

BUG FIXES:

• orchestrator/cattle: reload container's config before reading logs

1.0.0-alpha6

Changelog

BREAKING CHANGES:

engine/rclone: Don't format URL in rclone engine. Please use the standard rclone URL.

IMPROVEMENTS:

engine/restic: Check last backup date from bucket

BUG FIXES:

orchestrator/cattle: Improve log reading
orchestrator/cattle: Prevent failure when volume is mounted on stopped container
orchestrator/kubernetes: Fix current pod detection when using --all-namespaces

1.0.0-alpha5

Changelog

NOTES:

As we now forward environment variables from manager to workers, we plan to
remove the AWS and Swift options in the next release.
If you are using the Duplicity engine to backup on Swift, you'll probably have
to change the environment variables you pass to use the more standard OS__*
environment variables.

IMPROVEMENTS:

• core: Add template for OpenShift
• providers: Rewrite to make it more flexible
• core: Forward manager environment variables to worker
• core: Change the behaviour of GetMountedVolumes() and rename it to GetContainersMountingVolume()

BUG FIXES:

• engine/restic: Check errors when retrieving snapshots
• orchestrator/docker: Don't allocate pseudo-tty to worker so that stdout and stderr are not merged
• orchestrator/cattle: Increase client timeout
• orchestrator/cattle: Improve log reading

v1.0.0-alpha4

Changelog

BREAKING CHANGES:

• orchestrator/kubernetes: backup path now contains namespace instead of node name

IMPROVEMENTS:

• core: Add whitelist option to whitelist volumes
• orchestrator/kubernetes: Add --k8s-all-namespaces option to backup accross all namespaces
• engine/restic: Add support for swift auth v3
• engine/rclone: Add support for swift auth v3

BUG FIXES:

• core: Fix verification skipping if IsCheckScheduled is false
• orchestrator/kubernetes: Check for empty container statuses in Launchcontainer
• engine/restic: Prevent backup from silently failing

v1.0.0-alpha3

Release v1.0.0-alpha3

v1.0.0-alpha2

Release v1.0.0-alpha2

v1.0.0-alpha1

Release v1.0.0-alpha1