Nodetool repair

Repair is an offline process which synchronizes the data between nodes.

You should run the nodetool repair command regularly. If you delete data frequently, it should be more often than the value of gc_grace_seconds (by default: 10 days), for example, every week.

Scylla nodetool repair command supports the following options

  • -dc --in-dc execute the repair on the nodes in the specified named datacenter, the datacenter must be a local datacenter

    For example:

    nodetool repair -dc US_DC
    nodetool repair --in-dc US_DC
    
  • -et --end-token execute the repair on the node/s starting with the first token and ending with the specified UUID token

    For example:

    nodetool repair -et 90874935784
    nodetool repair --end-token 90874935784
    
  • -hosts --in-hosts execute the repair on a specific hosts

    For example:

    nodetool repair -hosts 8d5ed9f477644dbdbad843fddce94b7c
    nodetool repair --in-hosts 8d5ed9f477644dbdbad843fddce94b7c
    
  • -local --in-local-dc execute the repair on the nodes in the local datacenter only

    For example:

    nodetool repair -local
    nodetool repair --in-local-dc
    
  • -pr --partitioner-range execute the repair only on the first range returned by the partitioner

    For example:

    nodetool repair -pr
    nodetool repair --partitioner-range
    
  • -st --start-token execute the repair on a range of node/s starting with this token

    For example:

    nodetool repair -st 10474535988
    nodetool repair --start-token
    
  • keyspace execute the repair on a specific keyspace. The default is all of the keyspaces

    For example:

    nodetool repair <my_keyspace>
    
  • table execute the repair on a specific table or a list of space-delimited table names. The default is all of the tables.

    For example:

    nodetool repair <my_keyspace> <my_table>
    

Nodetool Reference