Files
madbase/docker-compose.pillar-database.yml
Vlad Durnea 38cab8c246
Some checks failed
CI/CD Pipeline / lint (push) Successful in 3m45s
CI/CD Pipeline / integration-tests (push) Failing after 58s
CI/CD Pipeline / unit-tests (push) Failing after 1m2s
CI/CD Pipeline / e2e-tests (push) Has been skipped
CI/CD Pipeline / build (push) Has been skipped
Verify M2/M3 implementation, fix regressions against M0/M1
Regressions fixed:
- gateway/src/worker.rs: missing session_manager field in AuthState (M3 regression)
- gateway/src/main.rs: same missing field in monolithic gateway
- storage/src/handlers.rs: removed unused validate_role (now handled by RlsTransaction)

M2 Storage Pillar — verified complete:
- StorageBackend trait with full API (put/get/delete/copy/head/list/multipart)
- AwsS3Backend implementation with streaming get_object
- StorageMode enum (Cloud/SelfHosted) in Config
- All routes: CRUD buckets, CRUD objects, copy, move, sign, public URL, health
- Bucket constraints: file_size_limit + allowed_mime_types enforced on upload
- TUS resumable uploads with S3 multipart (5MB chunking)
- Image transforms run via spawn_blocking
- docker-compose.pillar-storage.yml, templates/storage-node.yaml
- Shared Docker network on all pillar compose files

M3 Auth Completeness — verified complete:
- POST /logout revokes refresh tokens + Redis sessions
- GET /settings returns provider availability
- POST /magiclink with hashed token storage
- DELETE /user soft-delete with token revocation
- Recovery flow accepts new password
- Email change requires re-verification via token
- OAuth callback redirects with fragment tokens
- MFA verify returns aal2 JWT with amr claims
- MFA challenge validates factor ownership
- SessionManager wired into login/logout
- GET /sessions returns active sessions
- Configurable ACCESS_TOKEN_LIFETIME
- Claims model extended with session_id, aal, amr

Tests: 62 passed, 0 failed, 11 ignored (external services)
Warnings: 0
Made-with: Cursor
2026-03-15 14:40:48 +02:00

58 lines
1.3 KiB
YAML

# MadBase - Pillar: DB / State
# High-Availability PostgreSQL Cluster (Autobase)
services:
etcd:
image: quay.io/coreos/etcd:v3.5.9
container_name: madbase_etcd
environment:
- ETCD_INITIAL_CLUSTER_TOKEN=madbase-autobase
- ETCD_INITIAL_CLUSTER_STATE=new
volumes:
- etcd_data:/etcd-data
restart: unless-stopped
patroni:
image: registry.gitlab.com/postgres-ai/postgresql-autobase/patroni:3.0.2
container_name: madbase_patroni
environment:
- PATRONI_SCOPE=madbase-cluster
- PATRONI_ETCD3_HOSTS=etcd:2379
- PATRONI_POSTGRESQL_PASSWORD=${POSTGRES_PASSWORD}
volumes:
- db_data:/var/lib/postgresql/data
depends_on:
- etcd
restart: unless-stopped
haproxy:
image: haproxy:2.8-alpine
container_name: madbase_haproxy
volumes:
- ./autobase-haproxy.cfg:/usr/local/etc/haproxy/haproxy.cfg:ro
ports:
- "5432:5433"
- "6379:6379"
depends_on:
- patroni
- redis
restart: unless-stopped
redis:
image: redis:7-alpine
container_name: madbase_redis
command: redis-server --appendonly yes
volumes:
- redis_data:/data
restart: unless-stopped
volumes:
etcd_data:
db_data:
redis_data:
networks:
default:
name: madbase
external: true