When to run anti-entropy repair

When is repair needed?

Run repair in these situations:

• To routinely maintain node health.
  Note: Even if deletions never occur, schedule regular repairs. Setting a column to null is a delete.
• To recover a node after a failure while bringing it back into the cluster.
• To update data on a node containing data that is not read frequently, and therefore does not get read repair.
• To update data on a node that has been down.
• To recover missing data or corrupted SSTables. To do this, you must run non-incremental repair.

Guidelines for running routine node repair include:

• Run incremental repair daily, run full repairs weekly to monthly. Monthly is generally sufficient, but run more frequently if warranted.
  Important: Full repair is useful for maintaining data integrity, even if deletions never occur.
• Use the parallel and partitioner range options, unless precluded by the scope of the repair.
• Run a full repair to eliminate anti-compaction. Anti-compaction is the process of splitting an SSTable into two SSTables, one with repaired data and one with non-repaired data. This has compaction strategy implications.
  Note: Before you can run incremental repair, you must set the repaired state of each SSTable. For instructions, see Migrating to incremental repairs.
• Run repair frequently enough that every node is repaired before reaching the time specified in the gc_grace_seconds setting. Deleted data is properly handled in the cluster if this requirement is met.
• Schedule routine node repair to minimize cluster disruption.
  • If possible, schedule repair operation for low-usage hours.
  • If possible, schedule repair operations on single nodes at a time.
• Increase the time value setting of gc_grace_seconds if data is seldom deleted or overwritten. For these tables, changing the setting will:
  • Minimizes impact to disk space.
  • Allow longer interval between repair operations.
• Mitigate heavy disk usage by configuring nodetool compaction throttling options (setcompactionthroughput and setcompactionthreshold) before running a repair.

Guidelines for running repair on a downed node:

• Do not use partitioner range, -pr.