OIDC: 403 error on callback

[Ezehkiel]

Bug Description

When trying to use an external SSO with OIDC I got a 403 error on the callback URL. I tried with a custom SSO and with Google OIDC -> same behaviours :(

I added logs in the code and I was able to get the error :

Error: did not find expected authorization request details in session, req.session["64b2967b-5c9b-4cc5-860f-d4e16733084d"] is undefined
server-1                |     at /app/node_modules/openid-client/lib/passport_strategy.js:132:13
server-1                |     at OIDCAuthStrategy.authenticate (/app/node_modules/openid-client/lib/passport_strategy.js:191:5)
server-1                |     at OIDCAuthStrategy.authenticate (/app/packages/twenty-server/dist/src/engine/core-modules/auth/strategies/oidc.auth.strategy.js:26:22)
server-1                |     at attempt (/app/node_modules/passport/lib/middleware/authenticate.js:378:16)
server-1                |     at authenticate (/app/node_modules/passport/lib/middleware/authenticate.js:379:7)
server-1                |     at /app/node_modules/@nestjs/passport/dist/auth.guard.js:97:3
server-1                |     at new Promise (<anonymous>)
server-1                |     at /app/node_modules/@nestjs/passport/dist/auth.guard.js:89:83
server-1                |     at OIDCAuthGuard.<anonymous> (/app/node_modules/@nestjs/passport/dist/auth.guard.js:50:36)
server-1                |     at Generator.next (<anonymous>)

I'm not sure but my investigation point that the line that produce the error is

twenty/packages/twenty-server/src/engine/core-modules/auth/guards/oidc-auth.guard.ts

Line 63 in edf0a18

return (await super.canActivate(context)) as boolean;

I tried this on version v0.35.0 with a fresh install with docker. Here are my config files:

TAG=v0.35.0

#PGUSER_SUPERUSER=postgres
PGPASSWORD_SUPERUSER=-0Q2yWrpWQs16E3lKuE5gEgpf2JbgT0l

#PG_DATABASE_HOST=db
#REDIS_URL=redis://redis:6379

SERVER_URL=https://crm.test.ch

# Use openssl rand -base64 32 for each secret
APP_SECRET=jow5n7BsjOhYA97hM+ElyT7UZjaOY9oadWaH2QIndkU=
SIGN_IN_PREFILLED=true
AUTH_SSO_ENABLED=true
ENTERPRISE_KEY=foo
LOG_LEVELS=verbose
LOGGER_IS_BUFFER_ENABLED=false

STORAGE_TYPE=s3
STORAGE_S3_REGION=us-east-1
STORAGE_S3_NAME=twenty
STORAGE_S3_ENDPOINT=https://s3.test.ch
STORAGE_S3_ACCESS_KEY_ID=twenty_user
STORAGE_S3_SECRET_ACCESS_KEY=secret-key

name: twenty

services:
 change-vol-ownership:
   image: ubuntu
   user: root
   volumes:
     - server-local-data:/tmp/server-local-data
     - docker-data:/tmp/docker-data
   command: >
     bash -c "
     chown -R 1000:1000 /tmp/server-local-data
     && chown -R 1000:1000 /tmp/docker-data"

 server:
   image: twentycrm/twenty:${TAG}
   volumes:
     - server-local-data:/app/packages/twenty-server/${STORAGE_LOCAL_PATH:-.local-storage}
     - docker-data:/app/docker-data
   ports:
     - "80:3000"
   environment:
     PORT: 3000
     PG_DATABASE_URL: postgres://${PGUSER_SUPERUSER:-postgres}:${PGPASSWORD_SUPERUSER}@${PG_DATABASE_HOST:-db:5432}/default
     SERVER_URL: ${SERVER_URL}
     REDIS_URL: ${REDIS_URL:-redis://redis:6379}

     SIGN_IN_PREFILLED: ${SIGN_IN_PREFILLED}
     STORAGE_TYPE: ${STORAGE_TYPE}
     STORAGE_S3_REGION: ${STORAGE_S3_REGION}
     STORAGE_S3_NAME: ${STORAGE_S3_NAME}
     STORAGE_S3_ENDPOINT: ${STORAGE_S3_ENDPOINT}
     STORAGE_S3_ACCESS_KEY_ID: ${STORAGE_S3_ACCESS_KEY_ID}
     STORAGE_S3_SECRET_ACCESS_KEY: ${STORAGE_S3_SECRET_ACCESS_KEY}
     AUTH_SSO_ENABLED: ${AUTH_SSO_ENABLED}
     ENTERPRISE_KEY: ${ENTERPRISE_KEY}
     APP_SECRET: ${APP_SECRET}

     SENTRY_DSN: ${SENTRY_DSN}
     LOG_LEVELS: ${LOG_LEVELS}
     LOGGER_IS_BUFFER_ENABLED: ${LOGGER_IS_BUFFER_ENABLED}
     DEBUG_MODE: "true"
   sysctls:
     net.ipv6.conf.all.disable_ipv6: 1
   depends_on:
     change-vol-ownership:
       condition: service_completed_successfully
     db:
       condition: service_healthy
   healthcheck:
     test: curl --fail http://localhost:3000/healthz
     interval: 5s
     timeout: 5s
     retries: 10
   restart: always

 worker:
   image: twentycrm/twenty:${TAG}
   command: ["yarn", "worker:prod"]
   environment:
     PG_DATABASE_URL: postgres://${PGUSER_SUPERUSER:-postgres}:${PGPASSWORD_SUPERUSER}@${PG_DATABASE_HOST:-db:5432}/default
     SERVER_URL: ${SERVER_URL}
     REDIS_URL: ${REDIS_URL:-redis://redis:6379}
     DISABLE_DB_MIGRATIONS: "true" # it already runs on the server

     STORAGE_TYPE: ${STORAGE_TYPE}
     STORAGE_S3_REGION: ${STORAGE_S3_REGION}
     STORAGE_S3_NAME: ${STORAGE_S3_NAME}
     STORAGE_S3_ENDPOINT: ${STORAGE_S3_ENDPOINT}
     STORAGE_S3_ACCESS_KEY_ID: ${STORAGE_S3_ENDPOINT}
     STORAGE_S3_SECRET_ACCESS_KEY: ${STORAGE_S3_SECRET_ACCESS_KEY}
     AUTH_SSO_ENABLED: ${AUTH_SSO_ENABLED}
     ENTERPRISE_KEY: ${ENTERPRISE_KEY}
     APP_SECRET: ${APP_SECRET}

     SENTRY_DSN: ${SENTRY_DSN}
     LOG_LEVELS: ${LOG_LEVELS}
     LOGGER_IS_BUFFER_ENABLED: ${LOGGER_IS_BUFFER_ENABLED}
     DEBUG_MODE: "true"
   sysctls:
     net.ipv6.conf.all.disable_ipv6: 1
   depends_on:
     db:
       condition: service_healthy
     server:
       condition: service_healthy
   restart: always

 db:
   image: twentycrm/twenty-postgres-spilo:${TAG}
   volumes:
     - db-data:/home/postgres/pgdata
   environment:
     PGUSER_SUPERUSER: ${PGUSER_SUPERUSER:-postgres}
     PGPASSWORD_SUPERUSER: ${PGPASSWORD_SUPERUSER:-postgres}
     ALLOW_NOSSL: "true"
     SPILO_PROVIDER: "local"
   healthcheck:
     test: pg_isready -U ${PGUSER_SUPERUSER:-postgres} -h localhost -d postgres
     interval: 5s
     timeout: 5s
     retries: 10
   restart: always

 redis:
   image: redis
   restart: always

volumes:
 docker-data:
 db-data:
 server-local-data:

Expected behavior

The expected behavior is that the user can connect himself to the application.

[AMoreaux]

Hi @Ezehkiel

I see that your config doesn't enable Redis. Is it expected?

Could you try with redis enable:

REDIS_URL=redis://redis:6379

[Ezehkiel]

Hi @AMoreaux,

This is expected, because in the docker compose there is the default value which is the same. If I display the variable inside the container, the endpoint is correct, cf :

But I also tried uncommenting it in the .env file, unfortunately, no change.

[AMoreaux]

I can't reproduce the issue. Are you available for a quick call to sharescreen with discord?

[Ezehkiel]

Thanks, I write you on Discord to setup a call :)

[FelixMalfait]

Closing since no update (maybe @Ezehkiel you ended up finding a solution yourself! feel free to share here for others). Thanks!