25.8.16 Stable: Add guard for local data lakes

[zvonand]

Changelog category (leave one):

• Critical Bug Fix (crash, data loss, RBAC) or LOGICAL_ERROR

Changelog entry (a user-readable short description of the changes that goes to CHANGELOG.md):

Add setting allow_local_data_lakes. Off by default.

CI/CD Options

Exclude tests:

• Fast test
• Integration Tests
• Stateless tests
• Stateful tests
• Performance tests
• All with ASAN
• All with TSAN
• All with MSAN
• All with UBSAN
• All with Coverage
• All with Aarch64
• All Regression
• Disable CI Cache

Regression jobs to run:

• Fast suites (mostly <1h)
• Aggregate Functions (2h)
• Alter (1.5h)
• Benchmark (30m)
• ClickHouse Keeper (1h)
• Iceberg (2h)
• LDAP (1h)
• Parquet (1.5h)
• RBAC (1.5h)
• SSL Server (1h)
• S3 (2h)
• Tiered Storage (2h)

[github-actions]

Workflow [PR], commit [9a74a55]

[alsugiliazova]

Audit Report for PR #1530

Confirmed Defects

Low: Incorrect previous_value in SettingsChangesHistory for allow_local_data_lakes

• Impact: Users with compatibility set to a version older than 25.8.16.10002 will have local data lakes remain disabled, even though in those older versions the feature was implicitly allowed (no guard existed).
• Anchor: src/Core/SettingsChangesHistory.cpp / addSettingsChanges(..., "25.8.16.10002", {...})
• Trigger: Set SET compatibility = '25.8.13.10001' (or any version < 25.8.16.10002) and attempt CREATE TABLE ... ENGINE = IcebergLocal(...) or icebergLocal(...).
• Why defect: applyCompatibilitySetting() iterates versions higher than the compatibility version and sets each affected setting to previous_value. For allow_local_data_lakes, previous_value was set to false. Before this PR, local data lakes had no guard—they were implicitly allowed. The correct previous_value for backward compatibility is true.
• Fix direction: Change previous_value to true: {"allow_local_data_lakes", true, false, "..."}.
• Regression test direction: Add test with compatibility set to pre-25.8.16.10002 version; verify IcebergLocal/deltaLakeLocal work without explicit allow_local_data_lakes.
• Subsystem: Core/Settings compatibility.
• Blast radius: Users relying on compatibility for backward compatibility.

Code evidence:

// src/Core/SettingsChangesHistory.cpp:41-43
addSettingsChanges(settings_changes_history, "25.8.16.10002",
{
   {"allow_local_data_lakes", false, false, "New setting to guard local data lake engines and table functions"},
});

[alsugiliazova]

PR #1530 CI Verification Report

CI Results: 224 success, 2 failures, 2 errors, 21 BROKEN

Failures (2)

#	Suite	Error	Related to PR	Flakiness (cidb, 90d)
1	Grype Scan (clickhouse-server altinitystable-alpine)	1 high/critical CVE (CVE-2026-2673)	No — container image security scan	N/A
2	Grype Scan (clickhouse-server altinitytest-alpine)	1 high/critical CVE (CVE-2026-2673)	No — container image security scan	N/A

Errors (2)

#	Suite	Error	Related to PR
1	Regression aarch64 clickhouse_keeper_no_ssl_2	Job did not complete	No — infrastructure
2	Regression aarch64 session_timezone	Job did not complete	No — infrastructure

BROKEN Tests (21) — All Known/Flaky, None Related to PR

#	Suite	Test	Annotation	Flakiness (cidb, 90d)
1-5	Integration (amd_asan 1/6, amd_binary 1/5, amd_tsan 1/6, arm_binary 1/4, arm_binary 4/4)	test_backup_restore_on_cluster/test_cancel_backup.py::test_shutdown_cancels_backup	INVESTIGATE - NETLINK_ERROR	29 failures / 9 days
6-9	Integration (amd_asan 6/6, amd_binary 2/5, amd_tsan 6/6, arm_binary 1/4)	test_backup_restore_on_cluster/test_different_versions.py::test_different_versions	INVESTIGATE - NETLINK_ERROR	118 failures / 25 days
10	Integration (amd_tsan, 3/6)	test_overcommit_tracker/test.py::test_user_overcommit	INVESTIGATE - AssertionError on tsan	13 failures / 3 days
11-12	Integration (amd_binary 1/5, arm_binary 4/4)	test_storage_s3_queue/test_0.py::test_streaming_to_many_views[ordered]	KNOWN - Unstable upstream	85 failures / 19 days
13	Integration (arm_binary 4/4)	test_dirty_pages_force_purge/test.py::test_dirty_pages_force_purge	KNOWN: #1369	7 failures / 2 days
14-19	Stateless (amd_asan 2/2, amd_debug async, amd_ubsan par)	03441_deltalake_clickhouse_public_datasets / 03441_deltalake_clickhouse_virtual_columns	INVESTIGATE - S3 sometimes unreachable	3 failures / 3 days
20	Stateless (amd_asan seq, amd_ubsan seq)	02700_s3_part_INT_MAX	KNOWN - Memory limit exceeded	Known
21	Stateless (amd_msan, seq, 1/2)	00024_random_counters	INVESTIGATE - random timeout	6 failures / 5 days

Regression Tests: 128 passed, 2 errors

All 128 regression suites passed (both Release and Aarch64), including:

• Iceberg (1, 2) — Aarch64 and Release — all passed
• AggregateFunctions (1, 2, 3) — all passed
• S3 (minio, aws_s3, gcs, azure) — all passed
• S3Export (part, partition) — all passed
• Parquet / ParquetS3 — all passed
• TieredStorage (local, minio, s3amazon, s3gcs) — all passed
• LDAP (authentication, external_user_directory, role_mapping) — all passed
• RBAC (1, 2, 3) — all passed
• SSLServer (1, 2, 3) — all passed
• ClickHouseKeeper — all passed (except aarch64 no_ssl_2: job did not complete)
• Alter (move, attach, replace) — all passed
• Common suites (swarms, engines, functions, data_types, dns, kafka, kerberos, etc.) — all passed

2 regression errors were infrastructure issues (jobs did not complete), not test failures.

Conclusion

PR #1530 is verified.

[mkmkme]

Antalya 25.8 is covered by #1531