- Raw devices Implement
- Manage volume group that user created
- Data Migration and Backup
- Nodes and devices selector
- Enrich operator setting
- Operator SDK framework
- Metric optimization
- TopolvmCluster status optimization
- PVC rollback
Not only lvm, operator could orchestrate csi driver that support raw devices. we should implement csi driver.
volume group that created by user has available storage capacity, use could config CR TopolvmCluster and then operator could manage these volume groups.
Make cronjob that do regular backup
Enrich node and device selectors that makes deployment more flexible.
Now only support config kubelet root dir and enable discover devices, later we should add resource request and limit, log level etc.
Use operator sdk framework to refactor controller.
Collect status and capacity information about each component pouring into prometheus.
Add and optimize the progress and health status of the deployment process.
If node down, operator should make pvc rollback let pod rescheduled to another node.