New features

Redshift backups

Supported features

• Create backups
• Clean backups
• Share backups with account

Feature not supported

• Copy snapshot to a region
• Copy snapshot to another aws account

Bug fixes

• Serverless deployments are working again (with fixed crons)

Improvements

Added SNS doco

Improvements / bug fixes

• Entities with special characters are translated into backup names with only allowed alphanumeric characters and hyphens. Consecutive hyphens are squashed into single character.

• Backup names are including hash of entity name, avoiding name collision of entities with different resource ids, but same names

New features

• Backup data is being written to S3 bucket, including sharing info.
• Above sharing info from S3 is used to 'pull shared backups' - that is copy backups from originating account into account where shelvery is running. This feature has been implemented for all 4 currently supported engines (ebs, ec2ami, rds, rds_cluster)

In cases with > 100 rds snapshots in running account, there was an endless loop causing Lambda to timeout.

Bugs fixed

In case of shelvery backed being marked with shelvery:backup=false tag, cleanup process would break for rds and rds cluster backups.

Improvements

RDS Backups and RDS Cluster backups in RDS_COPY_AUTOMATED_SNAPSHOT mode - in case there is no automated snapshots available, rds or rds_cluster backup will fallback to RDS_CREATE_SNAPSHOT mode

Bugfixes

If there is no entity available at the time of backup cleanup - backup cleanup will fail. This has been fixed now, as it is not necessity for resource being backed up to be present within AWS account at the time of backup being cleaned up.

Bug fixes

This PR has changes to take in consideration value of shelvery:create_backups tag. Only resources tagged with tag value of 1, True, TRUE or true will get backed up / cleaned up.

Improvements

Single resource backups

by using configuration value shelvery_select_entity you can select single resource to be backed up, or it's backups to be cleaned up

Metadata - entity_id

shelvery:entity_id is added as tag with value of original resource to shelvery backups. This makes tracing back to original resources possible even when resource is deleted (e.g. you CloudTrail can be searched for this)

EC2 AMI Backups

Backing up EC2 instances as Amazon Machine Images is now supported. All of the EBS block devices will be included in AMI. AMIs are created with "no reboot" option.

Integration tests

Testing library has been moved from nose to pytest. Also, integration tests for EBS volume is added with tests for

• Sharing snapshots with another account
• Copying snapshot to another region (DR snapshots)
• Create snapshot
• Cleanup snapshot

Running in single thread

To ago around async invocations for sharing and regional copy operations within context of integration tests, SHELVERY_MONO_THREAD environment variable support has been added. This will make all operations synchronous, when running in CLI mode. Lambda mode, however, does not apply to this.

v0.3.0

• RDS Cluster Backups
• README updated

Features added

• Configure backup behaviour through resource tags (e.g. tag ebs volume with shelvery:config:shelvery_keep_daily_backups=7)

• RDS Cleanup in Lambda environment

• RDS Share in Lambda environment

• Multiple levels of configuration

• Added proper documentation to README.md