admin_repair_table

List of tables to query. An asterisk returns all tables. The user can provide a single element (which will be automatically promoted to a list internally) or a list.

ID of the type per table.

Optional parameters. Allowed keys are:

• repair_policy – Corrective action to take. Allowed values are:

  • delete_chunks – Deletes any corrupted chunks.

  • shrink_columns – Shrinks corrupted chunks to the shortest column.

  • replay_wal – Manually invokes write-ahead log (WAL) replay on the table.

  • alter_table – Reset columns modification after incomplete alter column.

  • resync_replicas – Repairs replicated tables whose record counts differ across the worker ranks and toms: the copy with the most records is kept, and every other copy is replaced with it. That copy may include records from a load that did not complete, or records that a delete did not remove. A table whose copies all have the same number of records is not repaired, even if their contents differ. Views cannot be named in input parameter table_names; a named table that is not replicated is reported in output parameter repair_status. An asterisk selects every replicated table. The repair runs only while no other requests are running. The repaired tables are listed in the repaired_replicated_tables entry of the response output parameter info map.

• verify_all – If false only table chunk data already known to be corrupted will be repaired. Otherwise the database will perform a full table scan to check for correctness. Allowed values are:

  • true

  • false

  The default value is ‘false’.

The default value is an empty dict ( ).

List of repaired tables.

List of repair status by table.

Additional information. Allowed keys are:

• repaired_replicated_tables – If repair_policy is resync_replicas, comma-separated list of the replicated tables that were repaired.

The default value is an empty dict ( ).