database {database_id} does not exist while handling command

Error message

database {database_id} does not exist while handling command {command:?}

What it means

While handling a command that needs a database's inflight graph info, the controller cannot find the database in its checkpoint state. The start of the affected job is failed via notify_start_failed because there is no database graph to apply the command to.

Source

Thrown at src/meta/src/barrier/checkpoint/control.rs:346

                    | Command::SourceChangeSplit(_)
                    | Command::Throttle { .. }
                    | Command::CreateSubscription { .. }
                    | Command::AlterSubscriptionRetention { .. }
                    | Command::ConnectorPropsChange(_)
                    | Command::Refresh { .. }
                    | Command::ListFinish { .. }
                    | Command::LoadFinish { .. }
                    | Command::ResetSource { .. }
                    | Command::ResumeBackfill { .. }
                    | Command::InjectSourceOffsets { .. } => {
                        if cfg!(debug_assertions) {
                            panic!(
                                "new database graph info can only be created for normal creating streaming job, but get command: {} {:?}",
                                database_id, command
                            )
                        } else {
                            warn!(%database_id, ?command, "database does not exist while handling the command");
                            notifier.notify_start_failed(anyhow!("database {database_id} does not exist while handling command {command:?}").into());
                            return Ok(());
                        }
                    }
                },
            };

            database.handle_new_barrier(
                Some((command, notifier)),
                checkpoint,
                span,
                partial_graph_manager,
                &self.hummock_version_stats,
                worker_nodes,
            )
        } else {
            let database = match self.databases.entry(database_id) {
                Entry::Occupied(entry) => entry.into_mut(),
                Entry::Vacant(_) => {

View on GitHub (pinned to 6469eb736d)

Solutions

Check whether the database was dropped concurrently; re-create it and re-issue the job.
Retry the CREATE statement after confirming the database exists.
Avoid issuing DROP DATABASE while creation jobs are in-flight in that database.

Example fix

-- before: racing drop
DROP DATABASE db1;
-- after: wait for pending creations
SHOW JOBS IN DATABASE db1; -- ensure none creating
DROP DATABASE db1;

Defensive patterns

Strategy: validation

Validate before calling

-- ensure db exists and has no dropping race before creating jobs
SELECT 1 FROM rw_databases WHERE database_id = <id>;

Prevention

Don't DROP DATABASE while creation jobs are in-flight in it.
Check SHOW JOBS IN DATABASE before issuing DDL there.
Retry failed CREATEs after confirming the database exists.

When it happens

Trigger: A create/reschedule command references database_id that has been dropped (or never registered) by the time handle_new_barrier processes it — e.g. DROP DATABASE raced with a pending CREATE job in that database.

Common situations: Dropping a database while streaming jobs are still being created in it; concurrent DDL racing barrier scheduling; stale command queued after database deletion.

Understand the failure class

Background: "Not found" and "does not exist" errors: why "Task not found", "No such folder", and "Can't find" fire when a lookup comes back empty — this error's family across 14 libraries.

Related errors

AI-assisted analysis of risingwavelabs/risingwave@6469eb736d (2026-09-11). Data as JSON: /api/errors/1fb1aebf584a23c8. Report an issue: GitHub.