apache/seatunnel · warning

Cannot drop replication slot '{}' because it's still in use

Error message

Cannot drop replication slot '{}' because it's still in use

What it means

dropReplicationSlot tried to drop a replication slot via pg_drop_replication_slot but the PostgreSQL backend reports SQLSTATE 55006 (OBJECT_IN_USE) because a session is still streaming from the slot. It retries ATTEMPTS times, then warns and returns false (slot not dropped).

Source

Thrown at seatunnel-connectors-v2/connector-cdc/connector-cdc-opengauss/src/main/java/io/debezium/connector/postgresql/connection/PostgresConnection.java:458

     *
     * @param slotName the name of the replication slot, may not be null
     * @return {@code true} if the slot was dropped, {@code false} otherwise
     */
    public boolean dropReplicationSlot(String slotName) {
        final int ATTEMPTS = 3;
        for (int i = 0; i < ATTEMPTS; i++) {
            try {
                execute("select pg_drop_replication_slot('" + slotName + "')");
                return true;
            } catch (SQLException e) {
                // slot is active
                if (PSQLState.OBJECT_IN_USE.getState().equals(e.getSQLState())) {
                    if (i < ATTEMPTS - 1) {
                        LOGGER.debug(
                                "Cannot drop replication slot '{}' because it's still in use",
                                slotName);
                    } else {
                        LOGGER.warn(
                                "Cannot drop replication slot '{}' because it's still in use",
                                slotName);
                        return false;
                    }
                } else if (PSQLState.UNDEFINED_OBJECT.getState().equals(e.getSQLState())) {
                    LOGGER.debug("Replication slot {} has already been dropped", slotName);
                    return false;
                } else {
                    LOGGER.error("Unexpected error while attempting to drop replication slot", e);
                    return false;
                }
            }
            try {
                Metronome.parker(Duration.ofSeconds(1), Clock.system()).pause();
            } catch (InterruptedException e) {
            }
        }
        return false;

View on GitHub (pinned to cf67b549a7)

Solutions

  1. Find and terminate the backend holding the slot: SELECT pg_terminate_backend(active_pid) FROM pg_replication_slots WHERE slot_name='<slot>'; then retry the drop.
  2. Close all other replication connections using the slot before dropping it.
  3. Wait for the streaming session to end (slot active=false in pg_replication_slots) and retry.
  4. If the slot must go regardless, run SELECT pg_drop_replication_slot('<slot>'); manually after terminating the holder.

Example fix

// before
dropReplicationSlot("dbz"); // returns false, slot in use
// after
// stmt.execute("SELECT pg_terminate_backend(active_pid) FROM pg_replication_slots WHERE slot_name='dbz' AND active");
// then dropReplicationSlot("dbz");
Defensive patterns

Strategy: try-catch

Validate before calling

SELECT active, active_pid FROM pg_replication_slots WHERE slot_name='<slot>'; -- only drop when active=false

Try / catch

try { boolean dropped = conn.dropReplicationSlot(slot); if (!dropped) { terminateHolderPid(slot); retryDrop(slot); } } catch (SQLException e) { /* handle */ }

Prevention

When it happens

Trigger: Calling connection.dropReplicationSlot(slotName) while another connection/thread holds an active replication stream on that slot; the last retry hits OBJECT_IN_USE again and returns false.

Common situations: Connector restart while the old task still streams; a orphaned walreceiver backend keeping the slot active; tests shutting down a connection while the replicator thread is alive; killing a job whose streaming connection hasn't been released server-side yet.

Understand the failure class

Background: 'Could not be found', 'does not exist', 'not found in database': the resource-not-found family when an ID, slug, key, or URI lookup comes back empty — this error's family across 20 libraries.

Related errors

AI-assisted analysis of apache/seatunnel@cf67b549a7 (2026-09-10). Data as JSON: /api/errors/ce0a710b60b250c6. Report an issue: GitHub.