Use a storage pipeline
Sometimes, before getting to the point where operations are fully consumed by the storage driver, it is useful to apply some filters, transformations or simple registrations. In order to keep the transport layer as much agnostic as possible from what the storage layer does, this blueprint proposes the creation of a pipeline that acts as 'middleware' for the storage layer.
The pipeline must be lightweight and return quickly and has to be mapped 1:1 to what the storage API exposes.
Blueprint information
- Status:
- Complete
- Approver:
- Flavio Percoco
- Priority:
- Medium
- Drafter:
- Flavio Percoco
- Direction:
- Approved
- Assignee:
- Flavio Percoco
- Definition:
- Approved
- Series goal:
- Accepted for icehouse
- Implementation:
- Implemented
- Milestone target:
- 2014.1
- Started by
- Flavio Percoco
- Completed by
- Flavio Percoco
Related branches
Related bugs
Sprints
Whiteboard
Gerrit topic: https:/
Addressed by: https:/
Implement common pipeline class
Addressed by: https:/
Use the pipeline for the storage layer
Gerrit topic: https:/
Addressed by: https:/
Setup storage pipeline in the boostrap instead of driver base
Addressed by: https:/
'Persist' __getattr__ results
Addressed by: https:/
Use stevedore instead of importutils
Addressed by: https:/
Return a consumer function instead of consuming
Gerrit topic: https:/