test(table-catalog): expand live conformance guide (#4277)

Co-authored-by: Henry Guo <marshawcoco@users.noreply.github.com>
This commit is contained in:
Henry Guo
2026-07-05 15:41:08 +08:00
committed by GitHub
parent 188ab2131d
commit 5f2359de45
5 changed files with 669 additions and 1 deletions
@@ -7,6 +7,7 @@ import argparse
import json
import re
import shlex
import urllib.parse
from collections import OrderedDict
from io import StringIO
from typing import Any
@@ -488,6 +489,463 @@ def databend_command(*, dsn: str) -> str:
return shell_join(["databend-sql", "--dsn", dsn, "-f", "/tmp/rustfs-s3tables-databend-stage.sql"])
def rest_api_prefix(rest_path: str) -> str:
prefix = normalized_rest_path(rest_path)
if not prefix.endswith("/v1"):
prefix = f"{prefix}/v1"
return prefix
def rest_path_segment(value: str) -> str:
return urllib.parse.quote(value, safe="")
def warehouse_catalog_path(*, warehouse: str, rest_path: str, suffix: str = "") -> str:
return f"{rest_api_prefix(rest_path)}/{rest_path_segment(warehouse)}{suffix}"
def table_catalog_path(*, warehouse: str, namespace: str, table: str, rest_path: str, suffix: str = "") -> str:
return (
f"{rest_api_prefix(rest_path)}/{rest_path_segment(warehouse)}"
f"/namespaces/{rest_path_segment(namespace)}/tables/{rest_path_segment(table)}{suffix}"
)
def live_conformance_evidence(
*,
warehouse: str,
namespace: str,
table: str,
rest_path: str,
metadata_location: str,
cleanup: bool,
) -> OrderedDict[str, Any]:
return OrderedDict(
[
("result", "operator-recorded"),
(
"required_run_metadata",
[
"rustfs_build",
"git_sha",
"catalog_backing",
"endpoint",
"warehouse",
"rest_path",
"namespace",
"table",
"metadata_location",
"client_name",
"client_version",
"run_timestamp_utc",
"operator",
"expected_status",
"observed_status",
"row_count",
"cleanup_result",
],
),
(
"result_table_template",
[
OrderedDict(
[
("client", "PyIceberg"),
("scenario", "create-append-reload-scan-direct-rest-probes"),
("expected_status", "pass"),
("expected_row_count", 2),
("claim_after_pass", "automated-smoke"),
]
),
OrderedDict(
[
("client", "Spark Iceberg REST catalog"),
("scenario", "create-namespace-create-table-append-refresh-count-cleanup"),
("expected_status", "pass"),
("expected_row_count", 2),
("claim_after_pass", "manual-live-verified"),
]
),
OrderedDict(
[
("client", "Trino Iceberg REST catalog"),
("scenario", "read-count-existing-table"),
("expected_status", "pass"),
("expected_row_count", 2),
("claim_after_pass", "manual-live-read-verified"),
("write_claim_after_pass", "not-claimed"),
]
),
OrderedDict(
[
("client", "DuckDB Iceberg"),
("scenario", "iceberg-scan-current-metadata-location"),
("expected_status", "pass"),
("expected_row_count", 2),
("claim_after_pass", "manual-live-read-verified"),
("write_claim_after_pass", "not-claimed"),
]
),
OrderedDict(
[
("client", "Databend"),
("scenario", "s3-stage-read-table-data-files"),
("expected_status", "pass"),
("claim_after_pass", "manual-live-s3-stage-verified"),
("iceberg_rest_catalog_claim_after_pass", "not-claimed"),
]
),
OrderedDict(
[
("client", "Snowflake Open Catalog / Iceberg integrations"),
("scenario", "operator-adapted-external-volume-reference"),
("expected_status", "operator-recorded"),
("claim_after_pass", "reference-only"),
("live_rustfs_interoperability_after_pass", "not-claimed"),
]
),
],
),
(
"promotion_rules",
[
"Keep PyIceberg as the only automated claim unless the run is executed by CI or a repeatable operator job.",
"Promote Spark only to manual-live-verified when the exact RustFS build, Spark version, Iceberg version, SQL output, and row_count are recorded.",
"Do not promote Trino or DuckDB write compatibility from read probes; write compatibility remains not-claimed.",
"Do not promote Snowflake or vendor catalog interoperability from a generated template without a repeatable live run.",
"Treat manual-live failures as compatibility findings and keep the previous public claim boundary.",
],
),
(
"target",
OrderedDict(
[
("warehouse", warehouse),
("namespace", namespace),
("table", table),
("rest_path", rest_path),
("metadata_location", metadata_location),
("cleanup", cleanup),
]
),
),
]
)
def production_operations_guide(
*,
endpoint: str,
warehouse: str,
namespace: str,
table: str,
rest_path: str,
) -> OrderedDict[str, Any]:
endpoint = normalized_endpoint(endpoint)
table_base = table_catalog_path(warehouse=warehouse, namespace=namespace, table=table, rest_path=rest_path)
pyiceberg_smoke_command = shell_join(
[
"python3",
"scripts/table-catalog/pyiceberg_smoke.py",
"--profile",
"rustfs",
"--endpoint",
endpoint,
"--bucket",
warehouse,
"--namespace",
namespace,
"--table",
table,
"--rest-path",
rest_path,
"--replace",
"--cleanup",
]
)
live_conformance_command = shell_join(
[
"python3",
"scripts/table-catalog/engine_compatibility.py",
"--endpoint",
endpoint,
"--warehouse",
warehouse,
"--namespace",
namespace,
"--table",
table,
"--rest-path",
rest_path,
"--print-live-conformance",
"--cleanup",
]
)
return OrderedDict(
[
("claim_boundary", "Iceberg REST Catalog S3 Tables implementation"),
(
"target",
OrderedDict(
[
("endpoint", endpoint),
("warehouse", warehouse),
("namespace", namespace),
("table", table),
("rest_path", rest_path),
]
),
),
(
"sections",
[
OrderedDict(
[
("name", "live-client-conformance"),
("objective", "Record repeatable client evidence before expanding engine compatibility claims."),
(
"commands",
[
pyiceberg_smoke_command,
live_conformance_command,
],
),
(
"required_evidence",
[
"RustFS build and git SHA",
"catalog backing mode",
"client name and version",
"generated command or SQL",
"expected and observed status",
"row_count or response status",
"current metadata location",
],
),
(
"pass_criteria",
[
"PyIceberg append/reload/scan returns row_count=2",
"Spark manual/live run returns row_count=2 before cleanup",
"read-only engines do not claim write compatibility",
],
),
(
"fail_closed_signals",
[
"missing run metadata",
"manual-live run fails",
"client writes are attempted through a read-only probe",
],
),
]
),
OrderedDict(
[
("name", "catalog-backing-cutover"),
("objective", "Verify durable backing readiness before selecting a non-object-backed catalog mode."),
("commands", [f"GET {warehouse_catalog_path(warehouse=warehouse, rest_path=rest_path, suffix='/catalog/migration')}"]),
(
"required_evidence",
[
"migration dry-run response",
"commit recovery blockers",
"idempotency index readiness",
"warehouse prefix index readiness",
"rollback configuration",
],
),
(
"pass_criteria",
[
"blockers is empty",
"recommended actions are completed",
"object-backed catalog backup exists before cutover",
],
),
(
"fail_closed_signals",
[
"non-empty blockers",
"recoverable commit gaps require repair",
"warehouse prefix index is missing or stale",
],
),
]
),
OrderedDict(
[
("name", "maintenance-operations"),
("objective", "Validate controlled maintenance without claiming a continuous in-process scheduler."),
(
"commands",
[
f"POST {table_base}/maintenance/metadata",
f"GET {table_base}/maintenance/jobs/{{job}}",
],
),
(
"required_evidence",
[
"dry-run report before delete or rewrite",
"current pointer re-read",
"audit-events for planning and worker transitions",
"quarantine status for failed jobs",
"safe-window configuration",
],
),
(
"pass_criteria",
[
"unreachable object deletion candidates pass safety-window checks",
"rewrite groups stay partition-local and sort-order-local",
"delete-file or row-level maintenance remains manual review",
],
),
(
"fail_closed_signals",
[
"stale maintenance plan",
"quarantine required",
"row-level delete file rewrite requested",
"missing referenced object",
],
),
]
),
OrderedDict(
[
("name", "recovery-and-disaster-rehearsal"),
("objective", "Exercise operator recovery paths without moving the table pointer unexpectedly."),
(
"commands",
[
f"GET {table_base}/catalog/diagnostics",
f"POST {table_base}/catalog/recovery",
"python3 scripts/table-catalog/failure_coverage.py --print-disaster-recovery-rehearsal",
],
),
(
"required_evidence",
[
"catalog export baseline",
"diagnostics recovery status",
"safe repair result",
"rollback or import response when used",
"post-recovery loadTable result",
],
),
(
"pass_criteria",
[
"safe repair does not move the current pointer",
"rollback/import uses normal catalog validation",
"post-recovery loadTable returns the expected metadata location",
],
),
(
"fail_closed_signals",
[
"manual-review diagnostics",
"stale rollback or import conflict",
"post-recovery data-plane policy failure",
],
),
]
),
OrderedDict(
[
("name", "permissions-and-credentials"),
("objective", "Verify table policy and credential vending cannot bypass table warehouse scope."),
(
"commands",
[
"python3 scripts/table-catalog/pyiceberg_smoke.py --profile rustfs-vended-credentials --replace --cleanup",
f"GET {table_base}/credentials",
],
),
(
"required_evidence",
[
"catalog principal permissions",
"vended credential expiration",
"warehouse prefix scope",
"inside-prefix S3 put/head/get/delete result",
"outside-prefix deny result",
],
),
(
"pass_criteria",
[
"default credentials response is empty unless vending is enabled",
"vended credentials work inside the exact table prefix",
"vended credentials are denied outside the exact table prefix",
],
),
(
"fail_closed_signals",
[
"long-lived storage secret returned by catalog",
"prefix mismatch",
"outside-prefix object access succeeds",
],
),
]
),
OrderedDict(
[
("name", "unsupported-claim-governance"),
("objective", "Keep release language aligned with verified RustFS behavior."),
(
"commands",
[
"python3 scripts/table-catalog/pyiceberg_smoke.py --print-unsupported-inventory",
"python3 scripts/table-catalog/engine_compatibility.py --print-engine-matrix",
],
),
(
"not_claimed",
[
"full AWS S3 Tables control-plane API parity",
"full MinIO AIStor private extension parity",
"full Cloudflare R2 Data Catalog interoperability",
"active-active multi-region table writes",
"multi-table transactions",
"built-in SQL query execution",
"Delta Lake or Hudi table format support",
"end-to-end SQL row-level DML validation",
],
),
(
"required_evidence",
[
"support matrix status label",
"unsupported inventory entry",
"live evidence before any promotion",
],
),
(
"pass_criteria",
[
"public claim wording matches support matrix status",
"reference profiles stay reference-only without live evidence",
],
),
(
"fail_closed_signals",
[
"vendor parity wording without recorded live evidence",
"manual probe documented as automated support",
],
),
]
),
],
),
]
)
def live_conformance_harness(
*,
endpoint: str,
@@ -600,6 +1058,17 @@ def live_conformance_harness(
]
),
),
(
"evidence",
live_conformance_evidence(
warehouse=warehouse,
namespace=namespace,
table=table,
rest_path=rest_path,
metadata_location=metadata_location,
cleanup=cleanup,
),
),
(
"clients",
[
@@ -731,6 +1200,7 @@ def parse_args(argv: list[str] | None = None) -> argparse.Namespace:
parser.add_argument("--cleanup", action="store_true")
parser.add_argument("--print-engine-matrix", action="store_true")
parser.add_argument("--print-live-conformance", action="store_true")
parser.add_argument("--print-operations-guide", action="store_true")
parser.add_argument("--print-spark-config", action="store_true")
parser.add_argument("--print-spark-sql", action="store_true")
return parser.parse_args(argv)
@@ -808,6 +1278,20 @@ def run(args: argparse.Namespace, output: StringIO | None = None) -> None:
output,
)
printed = True
if args.print_operations_guide:
print_json(
{
"production_operations_guide": production_operations_guide(
endpoint=args.endpoint,
warehouse=args.warehouse,
namespace=args.namespace,
table=args.table,
rest_path=args.rest_path or "/iceberg",
)
},
output,
)
printed = True
if args.print_spark_sql:
sql = spark_sql_smoke(
catalog_name=args.catalog_name,