Administrator Operations
This section documents operations required by administrators to maintain an Ozone system in a production environment.
📄️ Start and Stop
This guide describes how to start and stop an Ozone cluster, assuming it has already been configured and initialized.
📄️ Upgrade and Downgrade
Ozone supports non-rolling upgrades and downgrades, where all components are stopped first, and then restarted with the upgraded or downgraded versions.
🗃️ Decommissioning and Maintenance
3 items
🗃️ Disk Replacement
4 items
📄️ Container Replication Report
Ozone continuously monitors the health of containers to handle the loss of disks or nodes and re-replicates data to maintain a healthy replication count. The status of this replication can be viewed with the Container Report command which can be run as follows:
🗃️ Data Balancing
2 items
🗃️ S3 Multi-Tenancy
4 items
🗃️ Snapshots
2 items
🗃️ Observability
6 items
🗃️ Leader Transfer
2 items
📄️ Quota in Ozone
Client usage
🗃️ Tools
3 items
📄️ Trash
The trash feature in Ozone provides a way to recover files that have been accidentally deleted. When enabled, deleted files are moved to a trash directory instead of being permanently removed.
📄️ Dynamic Property Reload
Ozone supports dynamic reloading of certain configuration properties without restarting services. This enables operators to tune cluster behavior, adjust limits, and update settings in production without service disruption.
📄️ Cluster expansion
This guide covers common ways to grow capacity: new Datanodes, extra disks on existing Datanodes, and more memory (heap) for a Datanode. These operations are supported; Ozone picks up new nodes and storage through normal Datanode registration and heartbeat metadata—no separate “sync” step is required beyond correct configuration and restart where noted.