What causes "Cannot get part name from system.parts" in apache/seatunnel?

Confirm the table exists and the user can run `SELECT database, table, name FROM system.parts WHERE database='..' AND table='..'` in clickhouse-client.

apache/seatunnel · error · ClickhouseConnectorException

GET_PART_ERROR

GET_PART_ERROR

Error message

Cannot get part name from system.parts

What it means

ClickhouseProxy.getPartList queries system.parts to enumerate data parts of a table (per shard) and wraps any ClickHouseException thrown by that query in GET_PART_ERROR. It means the metadata query against system.parts failed, so the reader cannot determine which parts to scan.

Solutions

Confirm the table exists and the user can run `SELECT database, table, name FROM system.parts WHERE database='..' AND table='..'` in clickhouse-client.
Check ClickHouse server availability and network connectivity from the SeaTunnel worker nodes.
Grant the connector user access to system.parts (it is normally world-readable; a custom users.xml restriction may block it).
Retry the job — transient ClickHouseException (timeouts) is surfaced here; add connection/read timeout tuning if it recurs.

Defensive patterns

Strategy: retry

Validate before calling

// ensure the metadata query works before split enumeration
SELECT count(*) FROM system.parts WHERE database = 'db' AND table = 'tbl' AND active;

Try / catch

try {
    proxy.getPartList(database, table, shard, ...);
} catch (ClickhouseConnectorException e) {
    LOG.warn("system.parts query failed, retrying", e);
    // bounded retry with backoff, then fail
}

Prevention

Verify the ClickHouse user can query system.parts for the target table.
Keep ClickHouse connections healthy (timeouts, health checks) so enumeration doesn't hit dead nodes.
Confirm database/table names resolve identically on all shards of a Distributed table.

When it happens

Trigger: Calling getPartList(database, table, ...) when the system.parts query fails: table doesn't exist, system.parts is filtered to zero-access for the user, or the server is unreachable during split enumeration.

Common situations: Reading from a replicated/distributed table where the configured database/table resolves differently per shard; the ClickHouse user lacks grants on system.parts; transient network failure to ClickHouse during job startup.

Understand the failure class

Background: Database query failed: Internal Server Error 500s wrapping SQL, Prisma, and connection failures — what to check first — this error's family across 16 libraries.

Related errors

AI-assisted analysis of apache/seatunnel@cf67b549a7 (2026-09-10). Data as JSON: /api/errors/3c663303cd943dda. Report an issue: GitHub.

Appendix: source

Thrown at seatunnel-connectors-v2/connector-clickhouse/src/main/java/org/apache/seatunnel/connectors/seatunnel/clickhouse/util/ClickhouseProxy.java:471

        if (partitionList != null && !partitionList.isEmpty()) {
            StringJoiner joiner = new StringJoiner("', '", "('", "')");
            partitionList.forEach(joiner::add);

            sql += " and partition in " + joiner.toString();
        }

        sql += " group by name";

        log.debug("get part sql: {}", sql);

        try (ClickHouseResponse response = clickhouseRequest.query(sql).executeAndWait()) {
            Iterable<ClickHouseRecord> records = response.records();
            return StreamSupport.stream(records.spliterator(), false)
                    .map(r -> new ClickhousePart(r.getValue(0).asString(), database, table, shard))
                    .collect(Collectors.toList());
        } catch (ClickHouseException e) {
            throw new ClickhouseConnectorException(
                    ClickhouseConnectorErrorCode.GET_PART_ERROR,
                    "Cannot get part name from system.parts",
                    e);
        }
    }

    public List<SeaTunnelRow> batchFetchRecords(
            String sql, TablePath tablePath, SeaTunnelRowType seaTunnelRowType) {
        List<SeaTunnelRow> seaTunnelRowList = new ArrayList<>();
        log.debug("run query data sql: {}", sql);

        try (ClickHouseResponse response = clickhouseRequest.query(sql).executeAndWait()) {
            response.stream()
                    .forEach(
                            record -> {
                                SeaTunnelRow seaTunnelRow =
                                        ClickhouseUtil.convertToSeaTunnelRow(
                                                record, seaTunnelRowType, tablePath.getFullName());

View on GitHub (pinned to cf67b549a7)