Perform cluster maintenance

Overview

MicroCeph provides a simple and consistent workflow to support maintenance activity.

Before proceeding, please refer to the Cluster maintenance to understand its functionality and impact.

Enabling Cluster Maintenance

To review the action plan for enabling maintenance mode, run

microceph cluster maintenance enter <node> --dry-run

If you only want to verify if the node is ready for maintenance operations, run

microceph cluster maintenance enter <node> --check-only

By default, noout is set when entering maintenance mode. To disable noout to enable data migration during maintenance, run

microceph cluster maintenance enter <node> --set-noout=False

By default, OSDs on the node are not stopped during maintenance mode, To stop the OSD service on the node during maintenance, run

microceph cluster maintenance enter <node> --stop-osds

You can also forcibly bring a node into maintenance mode or ignore the safety checks if you know what you are doing, but it’s generally not recommended as it’s not guaranteed the node is ready for maintenance operations.

# Forcibly enter maintenance mode
microceph cluster maintenance enter <node> --force

# Ignore safety checks when entering maintenance mode
microceph cluster maintenance enter <node> --ignore-check

Disabling Cluster Maintenance

To review the action plan for disabling maintenance mode, run

microceph cluster maintenance exit <node> --dry-run

To bring a node out of maintenance, run

microceph cluster maintenance exit <node>