sqe/docker-compose.test.yml at main · schubergphilis/sqe · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
# docker-compose.test.yml — Lightweight test stack (2 containers, no DB/Keycloak)
# Ports are offset by +10000 from production defaults to avoid conflicts.
# Usage:
#   docker compose -f docker-compose.test.yml up -d
#   ./scripts/bootstrap-test.sh
#   source tests/.test-env && cargo test -p sqe-coordinator --test integration_test -- --ignored

services:
  polaris:
    image: apache/polaris:1.7.0
    environment:
      POLARIS_PERSISTENCE_TYPE: in-memory
      POLARIS_BOOTSTRAP_CREDENTIALS: "POLARIS,root,s3cr3t"
      POLARIS_PRODUCTION_READINESS_CHECKS_ENABLED: "false"
      QUARKUS_HTTP_PORT: 8181
      QUARKUS_LOG_LEVEL: INFO
      QUARKUS_LOG_CATEGORY__ORG_APACHE_POLARIS__LEVEL: WARN
      quarkus.otel.sdk.disabled: "true"
      # 1.5.0 realm resolution is stricter than 1.3.0: declare the realm,
      # make it the default, and do not require the realm header (clients
      # in this stack never send one).
      polaris.realm-context.realms: POLARIS
      polaris.realm-context.default-realm: POLARIS
      polaris.realm-context.require-header: "false"
      # Allow DROP VIEW/TABLE to purge metadata files from object storage.
      # 1.5.0 reads feature flags via polaris.features."NAME"; the old
      # POLARIS_FEATURES_DEFAULTS_* env spelling is no longer honored.
      polaris.features."DROP_WITH_PURGE_ENABLED": "true"
      # S3-compatible endpoints without an STS service (e.g. StorageGRID)
      # cannot mint subscoped credentials — CREATE TABLE fails with
      # "Failed to get subscoped credentials: Status Code 405". RustFS
      # answers AssumeRole, so the local default keeps subscoping on;
      # benchmark-test.sh exports POLARIS_SKIP_SUBSCOPING=true for
      # BENCH_WAREHOUSE=external.
      polaris.features."SKIP_CREDENTIAL_SUBSCOPING_INDIRECTION": "${POLARIS_SKIP_SUBSCOPING:-false}"
      # S3 credentials and endpoint for Polaris' own S3 operations
      # (metadata writes, drop-with-purge). Default: the RustFS container.
      # benchmark-test.sh exports the POLARIS_S3_* variables to point the
      # warehouse at an external S3 endpoint (BENCH_WAREHOUSE=external);
      # a plain `docker compose up -d` restores the RustFS defaults.
      AWS_REGION: ${POLARIS_S3_REGION:-us-east-1}
      AWS_ACCESS_KEY_ID: ${POLARIS_S3_ACCESS_KEY:-s3admin}
      AWS_SECRET_ACCESS_KEY: ${POLARIS_S3_SECRET_KEY:-s3admin}
      QUARKUS_S3_ENDPOINT_OVERRIDE: ${POLARIS_S3_ENDPOINT:-http://rustfs:9000}
      QUARKUS_S3_PATH_STYLE_ACCESS: "true"
      # SDK-level default endpoint. With credential subscoping skipped
      # (external mode), Polaris' metadata-write S3 client is built from
      # the SDK default chain and ignores both QUARKUS_S3_ENDPOINT_OVERRIDE
      # and the catalog's storage endpoint — without this it falls through
      # to real AWS. The SDK env cannot force path-style, so the external
      # endpoint must support virtual-hosted addressing.
      AWS_ENDPOINT_URL_S3: ${POLARIS_S3_ENDPOINT:-http://rustfs:9000}
    # Virtual-host-style resolution for an external warehouse served from
    # the docker HOST (e.g. a native MinIO for fair-throughput benches on
    # macOS, where the docker proxy caps container->host at ~250MB/s).
    # Polaris' SDK-default metadata client cannot be forced to path-style
    # (see AWS_ENDPOINT_URL_S3 note above), so `warehouse.<host>` must
    # resolve; host-gateway points it at the host, and the store maps the
    # Host header to the bucket (MinIO: set MINIO_DOMAIN).
    extra_hosts:
      - "warehouse.host.docker.internal:host-gateway"
    ports:
      - "18181:8181"
    healthcheck:
      test: ["CMD", "curl", "--fail", "http://localhost:8182/q/health"]
      interval: 5s
      timeout: 3s
      retries: 15
      start_period: 15s

  rustfs:
    image: rustfs/rustfs:latest
    environment:
      RUSTFS_ACCESS_KEY: s3admin
      RUSTFS_SECRET_KEY: s3admin
      RUSTFS_ADDRESS: ":9000"
      RUSTFS_VOLUMES: /data
    volumes:
      - rustfs-data:/data
    ports:
      - "19000:9000"

  # PostgreSQL backing store for the iceberg-catalog-sql JDBC catalog.
  # Used by the Phase L pg_jdbc_catalog test that exercises the vendored
  # `iceberg-catalog-sql` crate against a real Postgres instance. Port
  # 15432 is offset from the production default (5432) to avoid
  # collisions on dev boxes.
  postgres:
    # Pinned to 13 to match renovate.json's `allowedVersions: "<14"` for
    # postgres. 18+ stores data in a major-version-specific subdirectory, so an
    # existing volume from an older image is unreadable and the container exits 1
    # before pg_isready ever passes. That is what broke the Ranger quickstart on
    # 2026-08-07; the follow-up pinned ranger-db and left this service on 18,
    # where nobody noticed because `TYPE jdbc` was unimplemented and so nothing
    # ever started it. If you already have a volume built by 18, reset it with
    # `docker compose -f docker-compose.test.yml down -v`.
    image: postgres:13-alpine
    environment:
      POSTGRES_USER: iceberg
      POSTGRES_PASSWORD: iceberg
      POSTGRES_DB: iceberg_catalog
    ports:
      - "15432:5432"
    healthcheck:
      test: ["CMD-SHELL", "pg_isready -U iceberg -d iceberg_catalog"]
      interval: 5s
      timeout: 3s
      retries: 15
      start_period: 5s
    volumes:
      - postgres-data:/var/lib/postgresql/data

volumes:
  rustfs-data:
  postgres-data: