DISTRIBUTED_PARTITIONS

DATABASE_ID

The unique ID of the related database for a given partition.

ORDINAL

The ordinal number for a given partition, which decides the order in which data is distributed among partitions.

HOST

The hostname or IP address for the related node.

PORT

The port number of the related node.

ROLE

The role of the related node.

LOCKED

Whether a given partition is locked, as part of a partition level lock or higher.

NODE_ID

The ID of the related node.

IS_OFFLINE

Whether a given partition is offline.

SYNC_DESIRED

This is TRUE if a given partition instance uses synchronous replication. This matches the column IS_SYNC from information_schema.DISTRIBUTED_DATABASES of the corresponding database.

STATE

The current state of a given partition. This can be EMPTY (for online masters) or

ASYNC - data is asynchronously replicated from the master partition.

SYNC - data is synchronously replicated from the master partition.

READY - a transition state from async to sync replication used to guarantee that replicas are connected and blocking commits if metadata indicates them as sync.

UNRECOVERABLE - due t a bug in writing out log records, or a resource issue that replay fails to handle appropriately, for example, disk IO issue, or out of memory (OOM). No user action can be done to bring it back to any other state.

TERM

The term of the database, used to uniquely identify which node was the master of a specific database.

INSTANCE_ID

Unique ID of the partition instance.

IS_READ_REPLICA

Whether a given partition is part of a read replica or not.

SITE_ID

The ID for the workspace.

SAFE_TS

The lower bound of the pre-prepare timestamps of any unresolved 2PC their participants (reference and partition databases) participate or will participate in.

READ_REPLICA_STATE

This can be used to determine which instance of a partition on a workspace (where a database is mounted read-only), is actually being used for query execution. Relevant for SingleStoreDB Cloud.