Offline Cluster Maintenance Steps

{question}

How do I take a cluster offline for host maintenance?

{question}

{answer}

For online maintenance (cluster is available during host maintenance), please see the guide 'Taking Leaves Offline without Cluster Downtime' in the SingleStore documentation. This guide is for trading off cluster downtime for a smaller maintenance window and less complex operation.

1. For each database, run 'SNAPSHOT DATABASE <databasename>;' from the Master Aggregator.
2. For each database run '_SYNC_SNAPSHOT <databasename>;' from the Master Aggregator.
3. Run 'SET GLOBAL leaf_failure_detection=OFF;` from the Master Aggregator.
4. Stop all SingleStore nodes.
5. Perform your maintenance on all hosts.
6. Start all SingleStore nodes.
7. Once all nodes have started, run 'SET GLOBAL leaf_failure_detection=ON;' from the Master Aggregator.

{answer}