copy data between sandboxes
Copy a whole data directory from a sandbox to another.
It's especially useful in replication, to save cycles. Load data to the master, and then transfer the data directory to the slaves without using the binary log.
Blueprint information
- Status:
- Complete
- Approver:
- Giuseppe Maxia
- Priority:
- Medium
- Drafter:
- Giuseppe Maxia
- Direction:
- Approved
- Assignee:
- Sandbox-developers
- Definition:
- Approved
- Series goal:
- Accepted for mysql-sandbox-3
- Implementation:
-
Implemented
- Milestone target:
- None
- Started by
- Giuseppe Maxia
- Completed by
- Giuseppe Maxia
Related branches
Related bugs
Sprints
Whiteboard
The jumping off point is a generic cloning process that copies the data directory from one sandbox to another. Key operations include:
* stopping both sandboxes
* (should we purge stuff from the data directories here (e.g. logs, etc)?
* copying contents of data from source to target (excluding logfiles, etc).
* starting both sandboxes
Input for this process is simply: source_sandbox and target_sandbox
Once the generic script is called, we can create three additional wrapper scripts. The first one will clone masters to slaves in an individual sandbox, and the only input needed would be the sandbox directory. The second wrapper script will actually clone an entire sandbox setup. Here I'm thinking we use the source sandbox's configuration to actually create a new empty sandbox, and then call the generic clone process to copy the data in. With this approach, we don't need to worry about recreating or editing the hard coded scripts (start, stop, etc) because they are already created for us.
Currently implemented through sbtool
Work Items
Dependency tree
* Blueprints in grey have been implemented.
